Hugging Face's logo

datablations
/
lm1-4b2-84b-c4-repetitions

TensorBoard
Model card Files Metrics Community
Muennighoff commited on
Commit
862213b
•
1 Parent(s): 1f7b9a4

Add files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +0 -0
  2. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_0.json +1 -0
  3. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_1.json +1 -0
  4. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_2.json +1 -0
  5. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_3.json +1 -0
  6. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_4.json +1 -0
  7. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_5.json +1 -0
  8. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_0.json +1 -0
  9. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_1.json +1 -0
  10. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_2.json +1 -0
  11. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_3.json +1 -0
  12. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_4.json +1 -0
  13. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_5.json +1 -0
  14. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_0.json +1 -0
  15. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_1.json +1 -0
  16. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_2.json +1 -0
  17. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_3.json +1 -0
  18. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_4.json +1 -0
  19. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_5.json +1 -0
  20. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_0.json +1 -0
  21. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_1.json +1 -0
  22. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_2.json +1 -0
  23. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_3.json +1 -0
  24. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_4.json +1 -0
  25. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_5.json +1 -0
  26. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_0.json +1 -0
  27. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_1.json +1 -0
  28. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_2.json +1 -0
  29. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_3.json +1 -0
  30. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_4.json +1 -0
  31. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_5.json +1 -0
  32. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_0.json +1 -0
  33. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_1.json +1 -0
  34. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_2.json +1 -0
  35. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_3.json +1 -0
  36. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_4.json +1 -0
  37. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_5.json +1 -0
  38. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_0.json +1 -0
  39. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_1.json +1 -0
  40. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_2.json +1 -0
  41. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_3.json +1 -0
  42. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_4.json +1 -0
  43. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_5.json +1 -0
  44. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_0.json +1 -0
  45. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_1.json +1 -0
  46. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_2.json +1 -0
  47. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_3.json +1 -0
  48. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_4.json +1 -0
  49. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_5.json +1 -0
  50. 4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_tldr_en_0.json +1 -0
.gitattributes CHANGED
The diff for this file is too large to render. See raw diff
 
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.4070835356827751, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03514958095848397}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.0758536616906455, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015747064380670645}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.3264375465319237, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004888854445231445}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.11509298027342854, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002040147114373331}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.03493638633069714, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0009342574915112234}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.15766160622381195, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0033114573324024405}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.0532813862747049, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012579627803205211}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.07257604824526195, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0014483785678009685}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.31637706878833355, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004769735504597033}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.1105412242108245, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0019072286738988954}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.0714774644843108, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014699104009543759}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.307939556685913, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004520814685280998}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.10843545057843905, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0019083088150967664}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.41914858834195134, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.030279335876129}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07536633674836868, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001620641410096321}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.3290768382699901, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00481767508183653}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.11424698089656772, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001973221738343803}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.03540467062379218, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001074817084017668}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.16089821041540717, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0033011630774406127}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.05368591058094131, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012551880063213156}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.07231503158237214, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015163361416883465}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.3189205930522712, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004694857387684187}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.10991123942051419, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018557651460448018}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.07148579673935408, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0015357817111525064}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.3110112645350247, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00441643475943137}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.1082043807305256, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018480349337665876}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.4241874936612034, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03699728854949305}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07469786641233617, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015771153206732972}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.32891693541469197, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004751520151482175}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.11375522621692136, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019642936162507533}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.03462695918297652, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0009079391487918842}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.16210166248343671, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003411098262587952}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.05344291957030947, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001233885317072834}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.07161852924412807, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0014761860684115284}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.31797917629392425, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004598849198704314}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.10934081987838024, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018415550723111455}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.0711214967812572, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00149377360051449}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.3130870814045286, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004421211065212564}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.10823991385374933, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018380668821100924}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.3916994292697065, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02655023153261868}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07713695315738618, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018521617901133295}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.32641437991318506, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004583689746653368}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.11443103117633296, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019845366723218495}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.036319480745632425, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0012079439649413412}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.15985213856119682, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003223582265695079}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.05368996382308088, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012403567348119643}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.07333561421491072, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0017170202297610163}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.3129899723170166, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004403504671395443}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.10920281437234497, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018472429946517301}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.07332584684616669, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017553031622823821}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.31070372160179327, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004314602758278112}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.10882089385505, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018674885337082484}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.37875018794247045, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.024296780304434905}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07231813177556075, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015118246585830762}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.31870699434574523, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0046463072458484975}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.1102139471634063, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019620155943445507}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.033695317164630666, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0009163247572691914}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.15554105747469235, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003236397171744527}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.0515680827205002, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001213141047008391}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06901574750871842, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0013947803448898716}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.30638147232578594, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004472401624140904}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.1054595308766645, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018290764447497754}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06885871876103705, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001420315845279487}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.30338100654345734, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0043561963135752306}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.10492767242713451, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001836832016012516}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_PALM_prompt_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.3689406693649318, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.01833284872989782}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.0725733890131515, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0016541722828599028}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.31647681542290346, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004649887369574888}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.10942321553706275, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001960578009336271}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.0340511038621137, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001109810266658632}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.1539259113242296, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003295678214536681}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.05107734688924233, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00122669666906548}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06920155517233274, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015510532870385023}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.30372126675172995, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004472388579309833}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.10453686352175766, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018279045748061345}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.0688242602910231, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0015744817424633028}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.3000075619025587, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004328528641012373}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.1036182486745027, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018321897232056268}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.2320345566104918, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002028663485069812}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.15806405453574032, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0018950997131969357}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.17529221125446331, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0014916739627182348}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.01677596652301549, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.000700778712609751}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.012168726274182532, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0006308549540794899}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.012985384177633208, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0005793970465797221}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.1871404213284981, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001604901624172935}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.12845338233190606, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001541707987416157}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.14159539007628982, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0011656632831698861}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2002817812856504, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0018638576492756192}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.13376513099281687, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0015142202748567432}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.14922590343653938, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0011984205257227288}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 0.041197050082785285, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.001108996485999182}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.1368123990805212, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.003859738443155984}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.21915722251491454, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005434745495444146}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.1478464813817612, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0038320333818157508}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.03979808957820905, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0017989632072343414}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.07332816379948168, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0030403554888855404}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.04640292562526275, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0020116942927143034}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.10824512629470052, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.003088739020891906}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.18075899176594987, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0044836417615243115}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.1170703502765758, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0029311547998135696}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.12042587524943701, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.003511890506396}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.19133333683300907, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004812564968367625}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.1291123762138599, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0034020744059904736}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.8996196002528378, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.19349493429212591}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.17608939105537802, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004253817230115363}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.3431140264090531, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00534101919038467}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.19633951658374446, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003958953046041264}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.06908776827206395, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.002528705334523847}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.13897905823799864, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0037630298921426447}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.07757017845101091, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.002400954415731867}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.14034375265741814, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0033583359725794662}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.29506101479424657, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004712859932989569}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.1586443874692295, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.003057777683650459}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.15250105487675392, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0037920002252605505}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.294858148838543, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004765884466280505}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.1689535186370901, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0034887482192121964}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.640138834989728, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06878839244992584}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.1480732637246692, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004206315352922732}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.3705823844540417, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005255405787076392}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.16851267425125402, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.00370915891341574}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.05851115547615061, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0023290778302875307}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.15887512559363004, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0038897602570469245}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.06808437331115559, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0021900035041626056}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.12082840824613766, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0033519196264855337}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.33255889008853323, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004852038022387119}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.14033970117873198, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0028896013836388623}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.12767131894850914, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0037644778529259577}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.3162971688824102, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004747043761926457}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.14398542256516328, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0033027404123591544}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.4187050896119742, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.05234246731012316}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.11219617600997056, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.003512511634359904}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.38024888704829407, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005180394391296343}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.13586161882742837, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003197161076846914}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.042085938124191064, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0018230760751544277}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.1645327089113456, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003997411490344945}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.052457179235399276, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0017515669453131311}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.09344570812882523, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0027200533279541257}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.35242896432602544, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0049265850395284186}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.11671197771344118, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0024641462705413553}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.09539660575352989, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0030856580577620744}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.32391745612838724, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004709017013998289}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.11463457129527055, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002803475929481998}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.245079321161806, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02778718321398768}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_explicit-graph-description2_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.09393710373652846, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.003199269737822915}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.3745567023525177, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005019704359364239}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.11729721313967732, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0027320054268922134}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.03517936482889984, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0016983516245859785}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.1605210440862932, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0038664434520114016}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.045075512409701604, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0015508059027240881}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.08008998231575896, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00256681324622759}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.35287444686063746, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004875421748566501}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.10332377556822082, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0021608967117124093}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.08023373310406072, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00288649450260082}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.3198596952956621, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004543895911760099}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.09887761108872525, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002397235033540902}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.1514814811735028, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.041945649669857865}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.020794810389244665, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.008286245115526584}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.022922154005776907, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.00035921812588188365}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.21352555655696376, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0022208454109318375}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.04006653282921872, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0005481740262477606}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.0010806676437478898, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 5.572788223887115e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.013144891498511075, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0006988181195494956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.0019179536475281184, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 8.819595205548223e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.02287171287886853, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00035207258394197473}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.21329473631474377, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0022102531793857646}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.03999564658771847, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0005417492107414437}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.014803873787478817, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00023955003605054956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.14458682760689323, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0015738069244451358}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.0259194629518608, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0003466200097817442}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.4861728821422337, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03584212201959426}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.05062962614453291, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0013437525777807101}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.3417100308532278, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00435704773787356}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.08279157012164147, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0018413217717905348}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.015902434132343327, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0007863360597732104}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.10098499560418767, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0033802047604672408}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.025426464984268635, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0011153275941210136}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.046198311688407115, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0010451508171373679}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.3277232393217278, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0040337705566764045}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.07649332751100299, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0014727630258473052}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.04203232751325336, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0012323483368875934}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.2835715586404924, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004168417537419584}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.06849359236303296, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0016968890488419251}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.7628040722794971, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02211715710461336}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.05503857963517623, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0012530235158013246}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.4237722308516944, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004678442445263298}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.09191514123583507, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0015945801148090116}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.019614084687761036, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006958667373462588}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.1680913399724686, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003936935332174803}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.032591510976486694, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0009225376298957598}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.051275890765769695, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001008153954165397}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.40628782480488934, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0044353577159538015}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.08631782016595303, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001327845800016554}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.04580878529585366, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001122992125462262}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.3581492753310056, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004381030885989483}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.07633989045692885, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014222030587289858}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.8888663782890923, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.031568607067727}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.05690926034885604, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0011475096014499507}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.45426869216384025, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004516571311698728}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.09625702326801779, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0015524610950718182}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.021645434712929023, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006521179846718135}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.19508795473756288, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0040357210440640995}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.036709719893509046, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0009246274500291746}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.05330602414074714, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.000960960782770985}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.4341135138594583, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004317367652444747}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.09064023633752837, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0013324331225587393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.04733129735724454, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0010260711095464196}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.3853940847006117, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004214948856669242}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.08003764881477311, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0013775813454877939}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.945854666461263, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0402112381571833}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.05604585253845832, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0011427724679781285}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.4538828578884873, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0044575960643494774}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.09460944755353914, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0014786963748578273}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.021574156553869385, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.000666980759824044}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.19771338204023536, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004119461394324271}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.036402813498665906, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0009043444520209133}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.05270992641742948, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.000977850644744796}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.4347653663874296, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004302103689446801}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.08940502094865169, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012882532753422126}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.046717493551301156, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001010133723082569}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.38717142823870215, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004286251024406202}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.07885538264707831, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0012911242439370375}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_implicit-graph-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.9354058693381648, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02810525494577349}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.054228941021206276, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001096252527267709}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.44052473635855544, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004434086426797913}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.09189669042348231, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0014444352152093109}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.020624829786179254, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006718344531212511}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.18657022488728872, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003970246553779722}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.03482971179628577, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0009022373095447741}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.05120088679459415, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0009633103342396731}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.4212363318994364, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004211846253373217}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.0870670682221663, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012838750789689254}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.04543807296984024, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0010157838089103063}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.37673581998691646, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0042662278147285095}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.07691182427135196, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0013060342828319171}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.023111606817375823, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.00034433259453187245}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.199898162382514, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0022216437870279087}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.04023278195426544, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0005477579043898359}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.002348515359030586, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 7.292209026413432e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.02366089115265813, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0008284819548104561}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.004150191718708099, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00012739617704077432}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.0229905917432032, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00034282566300562767}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.1990997429912886, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0022213443978533944}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.04002572630665756, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0005453486432541972}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.01905255519601842, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0002660747474795132}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.16903333590221456, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0018194040988673658}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.033238272140701595, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00042068255632681374}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.011674876863475004, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0021009000372254386}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.08967550123251418, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002937976752458282}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.4773024647388192, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005382160167374159}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.12976799696821192, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0027894949987997152}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.03875399838383589, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0017591539560822005}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.21057258380768887, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004063079045737644}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.054485621293343514, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0017012540588488038}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.0790538192004279, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0023876721978287347}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.45016078797153286, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.005136353137741063}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.11667620039186845, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002240166240092561}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.0758745124201561, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002591817404665531}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.40795420296814416, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004919957414755}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.1093056100326863, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002463062122648188}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.9921366651933168, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06271540865538455}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.09279613971509679, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002133744119404951}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.5489640435072753, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004743143133672053}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.14521489817363703, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002414506324594877}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.03909773513534565, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0011574465963972958}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.2581685386804721, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004193436933392039}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.061645677874947354, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0013858346723081972}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.08058580758928455, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001676284644699535}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.5047433145790685, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004508210716996378}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.1276872860164771, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018563839053403013}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.07756422907254948, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0018428480505735482}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.46809438833435923, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004457544768356877}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.12134285926676484, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0020676862675326735}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.2825526231821482, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.043492144074173136}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.09618625233754133, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002582113721152913}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.5457125057626717, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0045898974012682685}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.14643243275737228, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0026100965175248907}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.04417002172231704, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0017784684644853198}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.26721249890626364, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004205423420143083}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.06571024213935271, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0016742580356473582}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.08292794661598844, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0020960952284175775}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.49763673601262126, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0043726175282994455}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.12785451973740822, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002033342221993515}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.08127713958483572, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002294208086828085}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.4695938897289019, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004353276805355962}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.1235235922343905, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0022715172117069057}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.408739047525639, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.037077565091005064}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.09479449002481802, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002655577330384621}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.5432614689390652, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0045982556155288136}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.14141133839063644, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0024329396682929907}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.04363505527515881, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001672653936113288}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.27112334525352516, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00422467374714849}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.06374220282296517, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0014814715876127712}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.08141857432754715, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0021601024343818197}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.4913866663151205, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004409556454898837}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.1229318925720696, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0019091259065232668}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.08007354294909853, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0023461776768967193}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.46756254108881956, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004390247412461436}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.11894159333044178, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0020882148965706148}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.3615130495425574, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03941780905887599}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_non-explicit-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.09830029665183404, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0032438344323854184}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.5307155773373864, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004576617366471351}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.13901031487174892, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0027114384521542325}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.04702658162841678, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0021412388855118754}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.2655552533910532, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004201121728916801}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.0639493149144395, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0017580675198847523}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.08605550281508571, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.002809323179439482}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.4835198358351258, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004434793924159799}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.12271761011397488, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0022754941590082543}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.08393751091304703, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0029130881830766323}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.4589032783950323, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004394339374968956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.11803135941981421, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0024007564296031226}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.343037351813773, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.036140936708617705}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.019449928935063972, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0002996123768501138}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.1629409626311148, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014178344731677658}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.033848667122813925, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0004785862714798914}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 3.7290701550733404e-05, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 9.057889706362649e-06}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.000247591268503305, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 6.608651546632237e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 6.345797512857661e-05, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 1.5382459284932663e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.019449928935063972, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0002996123768501138}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.1629409626311148, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0014178344731677658}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.033848667122813925, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0004785862714798914}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.013350740331407294, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0001939630721901288}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.11827840682573859, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0010744315351548378}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.02335480003154784, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00031241488270193515}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.0024661624004243305, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.00022679513212848655}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.08707890601958627, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0024828597057122404}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4696545657848939, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0050361060172909415}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.1299893322446928, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0025804647383953576}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.03503697621648224, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0014028562973121522}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.20358535015420007, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004005463978193232}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.051815361827752766, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001524719270466829}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.0774736609326366, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0019875056608878654}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.4430174064260006, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004789275467133504}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.11759668044891276, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0020751461605538037}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.07410190870582686, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002187569905436837}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.40495877524310353, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004615318859024589}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.11034796536631672, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0022796848420741974}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.9141145179175494, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06536296915571908}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.08092828074865113, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001906490261574064}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.508339922983901, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004811650108707499}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.1274257819886352, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002067479681389862}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.03434375738172082, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0011440228205307629}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.23872647813314588, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004162934850661181}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.0538074768484528, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012562150972743444}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.07273561966493279, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001561121717880125}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.47591647381990865, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0046462730133157}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.11571836456900784, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0016908336646337485}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.06920233060909893, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017246238008069412}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.44001472970441646, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004502443832581523}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.10863800930732682, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018630619874616906}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 1.2160721235507446, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.04337158036151172}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.08332513771516135, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0019868347121842745}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.5227234573351196, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004548834912859631}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.1318424172773736, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002194221638712534}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.0368091501321238, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0012212135534082967}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.25680330637418675, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004245573310206607}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.0579898457998029, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0013623575809253766}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.07491478363930333, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0016259914368083285}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.4884403183279412, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0044306286785567766}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.1196985480814696, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001792364415355123}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.07137645686823287, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017762733333957057}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.4552935323318038, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004331432435421035}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.11277885328608343, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0019446200008606618}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 1.2968562517794118, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03981886990665034}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.08350295350572431, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001964565788491711}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.5269669069519911, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004464761280028578}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.13136185565213176, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002061912807178742}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.036905855206776556, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001145818807419939}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.26098146346727347, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004211890254662535}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.05793797811823835, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012534985992537161}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.07557252948320996, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0016449289985662312}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.49370616371308285, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004327952531421013}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.12003948650766051, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0017167875473952968}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.07149122309163901, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017785526726022418}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.4585698429370505, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004258474803913088}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.11220846020003944, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018460781766260106}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 1.3609245244267043, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.033697352727562094}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-web_nlg_en_very-explicit-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.08171511695579445, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018633798688914489}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.5205014192182988, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00444667928382001}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.12912740120963084, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.00200364840452344}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.03547040912515195, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0010877669331809023}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.2544026510929024, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.004229382794688773}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.05600461944766409, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0012334741587230486}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.07419114114811591, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015341927152941845}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.4903982438780839, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004319214854158963}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.11838679216919511, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0016466394081195558}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.07010518510850142, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001733506016922793}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.4514049055251742, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00422208781632271}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.11026059508958445, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018150927429308838}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 1.3127171878921704, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.04798969717597349}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.08086960514171361, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0014525200909349487}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.14073285537633912, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0021365692846394106}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.09516029675770697, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0014796463176875243}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.008229815685974943, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0003995476871805908}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.014689368147720344, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0008211286470989261}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.009594517812957653, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00045393089210895286}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.07093712041741049, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0011411235398140263}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.12604423823582933, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0017629787569386458}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.08409008430092453, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0011584868305005939}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.07781962570247349, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0013794354759233953}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.13531803934376344, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0020069265089291905}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.09152903732984705, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0013950176929762666}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 0.4446304965171368, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02101564136831389}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.12036832068446836, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001698822978107527}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.21036263585148327, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002320909333489002}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.14245264302938512, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0016832867172118543}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.015115117361918744, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.000577145021228364}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.02662558452098901, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0010503502526558268}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.017846850141455827, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0006531290340899643}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.09229239279906276, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0011398229641967105}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.16800779726654005, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0017662341943095999}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.1106783481573006, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.00113630622144049}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.11299875946868077, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0015782836399768998}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.19806432139905789, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.002160572168760569}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.13385718956690607, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0015602587391599571}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 0.8653788220477945, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0447073485750703}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.12987828446896582, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0017668539132634244}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.22449329045901115, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002446761016968778}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.15317578509745602, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0017377869775542976}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.019205230811756614, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006259651415811169}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.034096443155707416, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.001170242516537603}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.022535640055881916, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00069644020248688}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.09796065026425818, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001176944853197895}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.1766738402223476, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0019113957567420726}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.11716221211305587, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0011775173863940828}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.12148938565288571, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0016327095456720253}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.21077828674873728, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0022836578166374076}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.1434435746637532, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0016022296713075052}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.0323682964004037, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.05728647371845087}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.11777387943190545, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0020113312278699025}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.19469078652969996, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0027977487029445204}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.1336180051181457, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001957815366131097}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.018842490102852782, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.000680153964280099}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.03249985364518616, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0011921354860818573}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.021648290209856712, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0007245383572720933}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.08919191590884211, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0014541958760042968}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.153037543777827, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002210331899287786}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.10208979953114408, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001397295719317153}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.10958122883328779, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0018686579393142751}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.1819882090629157, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0026135117445184627}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.12445447248036409, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018122356410507367}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.1833686716294687, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.07615919722768849}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.03995701958166195, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0016095372311748183}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.06674356961372938, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002448300622281182}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.04366405939122397, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0015831202808765945}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.006552251076514466, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00045575183717078306}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.012072979341302284, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0008431856424377719}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.007536808369783641, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0004892115817681631}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.03139542886892282, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012706755231911566}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.054115916065237676, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002006663515515118}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.03438650141175432, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012227168700738242}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.03739170639098872, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0015159196007358158}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.062356306725636794, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00228581975031369}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.040753188180590226, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014756445728831532}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 0.28624026182603163, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.029349438130954253}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_article_summary_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.007186612272605283, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0009791397918263772}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.009626580155215095, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0009843679259005512}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.006297564938231645, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0006539758063676466}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.0009639306536740955, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00016209042335021826}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.0018204921865520307, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0003123750046265983}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.0011309269927620334, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0001883904593636651}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.005946725242200601, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0008795298956796886}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.007926123528146993, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0008039901837259593}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.005007392411738194, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0004993418818895954}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.006742477351676616, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0009493223215183244}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.008831833968476688, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0009029356485515189}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.005783114895049256, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.000599975916196901}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.5436993670367395e-07, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 2.3789544446448377e-07}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.06660092740337692, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0010310436131496232}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.11598354826382899, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014500523897201271}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.07882898461240369, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001040524409407945}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0028614323156141855, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.000180186892026584}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.004436869555896244, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0002993853890744701}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.003243321779952968, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00020438291184088177}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.06084499114194029, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0008567327435936374}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.10872582938119582, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001328386237011712}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.07272695606243273, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0008850843973306463}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.06025198497698202, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0009166556048550033}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.10601353160224014, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00132199783950748}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.07152252322090602, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0009232099420132897}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.09735038671178359, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.01912399145191819}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.08855732672619102, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015064472119031337}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.14544347329432597, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002411071227170801}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.10178108004797332, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0015749327774327397}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.008560544006655627, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0004413067196946586}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.01616782557058256, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.000930039738517821}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.010181112623842817, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0005119910300898817}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.06881259230018068, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0010402326401547975}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.1164247420147934, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0018577911811822583}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.07977466300345593, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001103140871324476}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.08290521041610925, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0013971705074040997}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.13668730020343445, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0022409703680440853}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.09536226516262464, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014532751500060744}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.6029781033915351, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.028174541781078276}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.10757274244241427, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001866638618407923}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.1768593713292845, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0028614155741575843}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.12386118481423918, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019396226221506604}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.01811267415569796, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006924773828138044}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.03241684505960385, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.001268207503708144}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.02117387153026309, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0007530823627085244}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.08462993923914566, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00135467976486583}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.14336450100833226, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002264556242154818}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.09831165574662265, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0014131960658590383}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.09907287995201415, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017198819213882605}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.16367866039452192, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0026624060636886954}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.11426827856742555, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0017893139149704248}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 1.168914893531667, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.05316184069367659}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.10720171803251193, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002082996644638251}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.16612796673148053, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002954548192082436}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.11809803100929343, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0020464135073259295}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.02009963124741285, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0008011300876613331}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.03248360016391421, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0012256498842740874}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.022197025590925616, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0007547906063236964}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.08482141911566571, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015735036854485902}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.1351988789967234, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002389596247655131}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.09416415981113054, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0015409418875020468}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.09848261188127233, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0019260433293872472}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.15267739552040577, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0027171840590306546}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.10837367169734866, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001874380599850097}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 1.550517196878354, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06993149956571287}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.03740904754421892, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001610680414155072}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.057441900471446455, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0023292227400479332}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.03946196457385396, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001572740226370851}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.007405435782553709, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0004938171762244464}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.012822151487257908, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.000886391837086689}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.008171833643724272, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0005061088990249447}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.0302784538841258, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001265349446879213}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.047977766809024824, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001937859235680591}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.03213695990330082, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012354431038129371}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.034287485555940835, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014785361480277362}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.05268920626830129, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0021423352455123288}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.036104092106741856, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014394317299840206}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.2906818896887448, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.032241949265799034}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_rephrase_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.006416762662669792, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0008067399599762491}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.008596118967962398, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0009759171996390612}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.005997866632095373, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0006446489437867567}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0012325600300179128, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00023262166513716688}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.0019877130664642207, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0003765849024360566}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0012168332924537228, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00019190718073544674}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.005180574608491138, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0006638732889323238}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.00723418284441452, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.000838041683119617}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.004877643128103321, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0005141597079202806}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.00601329806116974, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0007735773924533153}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.007945295399230258, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0009128968590461367}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.005540816115674005, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0005975088748642658}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 2.3695690005957795e-08, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 5.162992091823081e-08}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.08473975013844745, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0012652837122476013}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.1329672875133205, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014302651043992514}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.09629596799033903, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011603461268498176}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.003906179417613626, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00019031811304692613}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.0055735013712897765, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00029292378757282304}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0042667329498244436, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00020313195756774255}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.07735309410625876, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001069934210773379}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.1244809192327779, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001333807275878821}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.08883637966635893, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0010087038286843665}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.08109267338305781, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0012068363440650514}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.12723188277215594, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001357597198069678}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.09211538989861727, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010998765766529236}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.02719689123536131, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0044363498052874435}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.07936181927044636, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0011762568246866275}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.1308334139584999, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014823829252056944}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.0918764247127699, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011260140101502995}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.00393290828598269, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0002040513153487916}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.005507726691318861, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00030125280804493035}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.00423567615381497, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00021018422233903274}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.07417084074707798, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001018585184337875}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.12499113633870057, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0013966793071261054}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.08666723799936435, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.000999850410378775}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.07610364515575635, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0011064264710643904}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.12615810768661415, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001412447688376772}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.08825782099755484, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010565687929762516}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.10510472666663716, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02342215944592973}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.07947361253543833, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001193005591889888}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.1327421000432318, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0015486342539799732}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.09256018146781847, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011555403499640406}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.004199424248706065, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0002403958914677049}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.006627430748129385, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00047327454697217267}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.004697153886380661, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00027487360812121926}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.07476922060726728, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001036726053715039}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.127478433032578, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0014601336227070283}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.08788495344807927, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001034393917882335}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.07575953247297404, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0011219051960697426}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.12703786617602278, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001464179232269078}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.08832933606271764, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010830718460803723}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.19993488449475255, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.04329139775558018}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.07037274511161584, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0014120231404269865}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.11052779587284725, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0017290812995824128}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.07737568279361932, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001265246486643532}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.00392300118223772, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0002557581635323497}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.005589835600869042, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.000423176072989138}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0040651974203171634, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00025930635468911277}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.06512620770805046, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00121951222912035}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.10533676686618007, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0016204045193858512}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07258759044318532, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001122423619227686}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.06672709213349835, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0013263129049748312}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.10524011607734472, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0016343300862625232}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.07343581428013704, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0011835280613330765}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.24585325834528793, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.03702715577574997}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.02381329967362824, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0011028927436461343}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.0375355725294583, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014975837973007148}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.025475522197387277, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0010381041342986902}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.0019534338228814068, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00025424799653585257}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.002574143785393473, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00033042674187292473}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0018254913452193152, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0002100606076864114}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.021656975285866457, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0009662420131766899}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.035213960582413724, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0013920460584087664}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.023474073198495527, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009278481009221991}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.022475087207169998, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0010391682891249803}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.03568293304710748, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001419767775872114}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.0240648893225021, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0009742444146235487}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.0725273475872256, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.01184587819436072}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_summarize_above_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.003599842790127908, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0004357453635530341}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.005450956260566027, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0006157647454877155}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.003773178583642327, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0004233214040303885}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.00024100397622410201, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 6.428724333491122e-05}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.0003662081649364038, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00010026927803212084}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.00026468573039586365, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 6.948326335135335e-05}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.003261903403975752, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0003814441870483113}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.005138441375666529, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0005722130168539618}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.0034927853884506757, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0003810695069644319}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.003394903398841389, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00041292623146590925}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.005123453693921271, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.000573270656957646}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.00354234805200471, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00039437507462346326}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 3.4432997292695364e-10, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 5.054319054078497e-10}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
4b284b12bc4/eval/agg.4b284b12bc4_GEM-wiki_lingua_en_tldr_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_precision": 0.0505257980847339, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0009316535384306269}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_recall": 0.07944574030730557, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0013436416711421135}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_fmeasure": 0.05739499438745971, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.000959404112224303}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_precision": 0.0025522430280765624, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00019188137819214202}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_recall": 0.003889530091650386, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0003502492712853139}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_fmeasure": 0.002874313185982406, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00022791640812011725}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_precision": 0.04641940622808299, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0007892141876172595}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_recall": 0.07403750267734954, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0011941764542527037}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_fmeasure": 0.053008425140295905, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0008201296289562219}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_precision": 0.04816018092226108, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0008696938078640268}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_recall": 0.07599978162074517, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0012632037879506343}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_fmeasure": 0.05476968826480647, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0008953065390907387}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "bleu": 0.14859459498800928, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.019538924284114197}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-4b2-84b-c4-repetitions/4b284b12bc4/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}