Datasets:

Tasks:
Other
Size:
n<1K
ArXiv:
License:
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 45 new columns ({'dclm:food_and_dining', 'dclm:entertainment', 's2pdf:politics', 's2pdf:health', 'dclm:fashion_and_beauty', 'dclm:health', 'dclm:literature', 's2pdf:religion', 's2pdf:travel', 'dclm:history_and_geography', 'dclm:travel_and_tourism', 's2pdf:education_jobs', 's2pdf:fashion_beauty', 'dclm:adult_content', 'dclm:software', 's2pdf:food_dining', 'dclm:politics', 's2pdf:transportation', 'dclm:science_math_and_technology', 'dclm:social_life', 'dclm:home_and_hobbies', 'dclm:art_and_design', 'dclm:religion', 's2pdf:games', 's2pdf:entertainment', 'dclm:education_and_jobs', 's2pdf:science_tech', 's2pdf:sports_fitness', 's2pdf:software_dev', 'dclm:software_development', 'dclm:sports_and_fitness', 's2pdf:finance_business', 'dclm:industrial', 'dclm:transportation', 's2pdf:crime_law', 's2pdf:home_hobbies', 'dclm:crime_and_law', 's2pdf:literature', 'dclm:electronics_and_hardware', 'dclm:finance_and_business', 's2pdf:adult', 's2pdf:software', 'dclm:games', 's2pdf:industrial', 's2pdf:art_design'}) and 110 missing columns ({'mmlu_human_aging:rc::olmes', 'mt_mbpp_v2fix:c', 'basic_skills_common_knowledge:rc::olmes', 'minerva_math_prealgebra::olmes', 'mmlu_high_school_geography:rc::olmes', 'mmlu_high_school_statistics:rc::olmes', 'mmlu_professional_medicine:rc::olmes', 'mmlu_college_mathematics:rc::olmes', 'mt_mbpp_v2fix:python', 'naturalqs:rc::gen2mc', 'mmlu_high_school_chemistry:rc::olmes', 'mmlu_formal_logic:rc::olmes', 'mmlu_world_religions:rc::olmes', 'mmlu_international_law:rc::olmes', 'mmlu_electrical_engineering:rc::olmes', 'mmlu_sociology:rc::olmes', 'mt_mbpp_v2fix:cpp', 'mmlu_elementary_mathematics:rc::olmes', 'mmlu_moral_scenarios:rc::olmes', 'mbpp:3shot::none', 'mmlu_nutrition:rc::olmes', 'lab_bench_dbqa', 'basic_skills:rc::olmes', 'mmlu_high_school_government_and_politics:rc::olmes', 'hellaswag:rc::olmes', 'coqa:rc::gen2mc', 'mt_mbpp_v2fix:haskell', 'mmlu_high_school_us_history:rc::olmes', 'mmlu_econometrics:rc::olmes', 'mmlu_us_foreign_policy:rc::olmes', 'sciq:rc::olmo3', 'medmcqa:rc::none', 'mmlu_prehistory:rc::olmes', 'mmlu_high_school_european_history:rc::olmes', 'mmlu_management:rc::olmes', 'socialiqa:rc::olmes', 'mmlu_computer_security:rc::olmes', 'mt_mbpp_v2fix:swift', 'mmlu_moral_disputes:rc::olmes', 'mt_mbpp_v2fix:rust', 'mmlu_astronomy:rc::olmes', 'minerva_math_precalculus::olmes', 'mmlu_college_medicine:rc::olmes', 'mmlu_anatomy:rc::olmes', 'mt_mbpp_v2fix:typescript', 'codex_humaneval:3shot::none', 'mmlu_medical_genetics:rc::olmes', 'minerva_math_counting_and_probability::o
...
_college_biology:rc::olmes', 'minerva_math_geometry::olmes', 'mmlu_global_facts:rc::olmes', 'mmlu_miscellaneous:rc::olmes', 'basic_skills_string_operations:rc::olmes', 'mt_mbpp_v2fix:php', 'mmlu_high_school_biology:rc::olmes', 'mmlu_human_sexuality:rc::olmes', 'winogrande:rc::olmes', 'mmlu_professional_law:rc::olmes', 'mt_mbpp_v2fix:java', 'mmlu_security_studies:rc::olmes', 'mt_mbpp_v2fix:bash', 'basic_skills_pattern:rc::olmes', 'mmlu_professional_accounting:rc::olmes', 'mt_mbpp_v2fix:ruby', 'mmlu_high_school_macroeconomics:rc::olmes', 'piqa:rc::olmes', 'lambada', 'medqa_en:rc::none', 'mmlu_logical_fallacies:rc::olmes', 'sciriff_yesno:rc::olmes', 'mmlu_high_school_computer_science:rc::olmes', 'basic_skills_arithmetic:rc::olmes', 'mmlu_marketing:rc::olmes', 'mmlu_jurisprudence:rc::olmes', 'arc_challenge:rc::olmes', 'mmlu_high_school_microeconomics:rc::olmes', 'mmlu_high_school_world_history:rc::olmes', 'qasper_yesno:rc::olmes', 'mmlu_public_relations:rc::olmes', 'mt_mbpp_v2fix:matlab', 'csqa:rc::olmes', 'mmlu_high_school_physics:rc::olmes', 'mt_mbpp_v2fix:r', 'squad:rc::gen2mc', 'minerva_math_number_theory::olmes', 'mmlu_business_ethics:rc::olmes', 'mmlu_philosophy:rc::olmes', 'mmlu_virology:rc::olmes', 'minerva_math_algebra::olmes', 'mmlu_conceptual_physics:rc::olmes', 'mmlu_machine_learning:rc::olmes', 'mmlu_college_computer_science:rc::olmes', 'mt_mbpp_v2fix:csharp', 'mt_mbpp_v2fix:scala', 'jeopardy:rc::gen2mc', 'drop:rc::gen2mc', 'minerva_math_intermediate_algebra::olmes'}).

This happened while the csv dataset builder was generating data using

hf://datasets/allenai/olmix/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv (at revision 33b58d580ba2b439fd9ede85a173188141586db3), [/tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/ratios.csv)]

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 675, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              run: string
              name: string
              index: int64
              dclm:adult_content: double
              dclm:art_and_design: double
              dclm:crime_and_law: double
              dclm:education_and_jobs: double
              dclm:electronics_and_hardware: double
              dclm:entertainment: double
              dclm:fashion_and_beauty: double
              dclm:finance_and_business: double
              dclm:food_and_dining: double
              dclm:games: double
              dclm:health: double
              dclm:history_and_geography: double
              dclm:home_and_hobbies: double
              dclm:industrial: double
              dclm:literature: double
              dclm:politics: double
              dclm:religion: double
              dclm:science_math_and_technology: double
              dclm:social_life: double
              dclm:software: double
              dclm:software_development: double
              dclm:sports_and_fitness: double
              dclm:transportation: double
              dclm:travel_and_tourism: double
              s2pdf:adult: double
              s2pdf:art_design: double
              s2pdf:crime_law: double
              s2pdf:education_jobs: double
              s2pdf:entertainment: double
              s2pdf:fashion_beauty: double
              s2pdf:finance_business: double
              s2pdf:food_dining: double
              s2pdf:games: double
              s2pdf:health: double
              s2pdf:home_hobbies: double
              s2pdf:industrial: double
              s2pdf:literature: double
              s2pdf:politics: double
              s2pdf:religion: double
              s2pdf:science_tech: double
              s2pdf:software: double
              s2pdf:software_dev: double
              s2pdf:sports_fitness: double
              s2pdf:transportation: double
              s2pdf:travel: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 6749
              to
              {'run': Value('string'), 'name': Value('string'), 'index': Value('int64'), 'arc_challenge:rc::olmes': Value('float64'), 'arc_easy:rc::olmes': Value('float64'), 'basic_skills:rc::olmes': Value('float64'), 'basic_skills_arithmetic:rc::olmes': Value('float64'), 'basic_skills_coding:rc::olmes': Value('float64'), 'basic_skills_common_knowledge:rc::olmes': Value('float64'), 'basic_skills_logical_reasoning:rc::olmes': Value('float64'), 'basic_skills_pattern:rc::olmes': Value('float64'), 'basic_skills_string_operations:rc::olmes': Value('float64'), 'codex_humaneval:3shot::none': Value('float64'), 'coqa:rc::gen2mc': Value('float64'), 'csqa:rc::olmes': Value('float64'), 'drop:rc::gen2mc': Value('float64'), 'hellaswag:rc::olmes': Value('float64'), 'jeopardy:rc::gen2mc': Value('float64'), 'lab_bench_dbqa': Value('float64'), 'lab_bench_protocolqa': Value('float64'), 'lambada': Value('float64'), 'mbpp:3shot::none': Value('float64'), 'medmcqa:rc::none': Value('float64'), 'medqa_en:rc::none': Value('float64'), 'minerva_math_algebra::olmes': Value('float64'), 'minerva_math_counting_and_probability::olmes': Value('float64'), 'minerva_math_geometry::olmes': Value('float64'), 'minerva_math_intermediate_algebra::olmes': Value('float64'), 'minerva_math_number_theory::olmes': Value('float64'), 'minerva_math_prealgebra::olmes': Value('float64'), 'minerva_math_precalculus::olmes': Value('float64'), 'mmlu_abstract_algebra:rc::olmes': Value('float64'), 'mmlu_anatomy:rc::olmes': Value('float64'), 'mmlu_
              ...
              loat64'), 'mmlu_professional_law:rc::olmes': Value('float64'), 'mmlu_professional_medicine:rc::olmes': Value('float64'), 'mmlu_professional_psychology:rc::olmes': Value('float64'), 'mmlu_public_relations:rc::olmes': Value('float64'), 'mmlu_security_studies:rc::olmes': Value('float64'), 'mmlu_sociology:rc::olmes': Value('float64'), 'mmlu_us_foreign_policy:rc::olmes': Value('float64'), 'mmlu_virology:rc::olmes': Value('float64'), 'mmlu_world_religions:rc::olmes': Value('float64'), 'mt_mbpp_v2fix:bash': Value('float64'), 'mt_mbpp_v2fix:c': Value('float64'), 'mt_mbpp_v2fix:cpp': Value('float64'), 'mt_mbpp_v2fix:csharp': Value('float64'), 'mt_mbpp_v2fix:go': Value('float64'), 'mt_mbpp_v2fix:haskell': Value('float64'), 'mt_mbpp_v2fix:java': Value('float64'), 'mt_mbpp_v2fix:javascript': Value('float64'), 'mt_mbpp_v2fix:matlab': Value('float64'), 'mt_mbpp_v2fix:php': Value('float64'), 'mt_mbpp_v2fix:python': Value('float64'), 'mt_mbpp_v2fix:r': Value('float64'), 'mt_mbpp_v2fix:ruby': Value('float64'), 'mt_mbpp_v2fix:rust': Value('float64'), 'mt_mbpp_v2fix:scala': Value('float64'), 'mt_mbpp_v2fix:swift': Value('float64'), 'mt_mbpp_v2fix:typescript': Value('float64'), 'naturalqs:rc::gen2mc': Value('float64'), 'piqa:rc::olmes': Value('float64'), 'qasper_yesno:rc::olmes': Value('float64'), 'sciq:rc::olmo3': Value('float64'), 'sciriff_yesno:rc::olmes': Value('float64'), 'socialiqa:rc::olmes': Value('float64'), 'squad:rc::gen2mc': Value('float64'), 'winogrande:rc::olmes': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1347, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 980, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 884, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 947, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1736, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1889, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 45 new columns ({'dclm:food_and_dining', 'dclm:entertainment', 's2pdf:politics', 's2pdf:health', 'dclm:fashion_and_beauty', 'dclm:health', 'dclm:literature', 's2pdf:religion', 's2pdf:travel', 'dclm:history_and_geography', 'dclm:travel_and_tourism', 's2pdf:education_jobs', 's2pdf:fashion_beauty', 'dclm:adult_content', 'dclm:software', 's2pdf:food_dining', 'dclm:politics', 's2pdf:transportation', 'dclm:science_math_and_technology', 'dclm:social_life', 'dclm:home_and_hobbies', 'dclm:art_and_design', 'dclm:religion', 's2pdf:games', 's2pdf:entertainment', 'dclm:education_and_jobs', 's2pdf:science_tech', 's2pdf:sports_fitness', 's2pdf:software_dev', 'dclm:software_development', 'dclm:sports_and_fitness', 's2pdf:finance_business', 'dclm:industrial', 'dclm:transportation', 's2pdf:crime_law', 's2pdf:home_hobbies', 'dclm:crime_and_law', 's2pdf:literature', 'dclm:electronics_and_hardware', 'dclm:finance_and_business', 's2pdf:adult', 's2pdf:software', 'dclm:games', 's2pdf:industrial', 's2pdf:art_design'}) and 110 missing columns ({'mmlu_human_aging:rc::olmes', 'mt_mbpp_v2fix:c', 'basic_skills_common_knowledge:rc::olmes', 'minerva_math_prealgebra::olmes', 'mmlu_high_school_geography:rc::olmes', 'mmlu_high_school_statistics:rc::olmes', 'mmlu_professional_medicine:rc::olmes', 'mmlu_college_mathematics:rc::olmes', 'mt_mbpp_v2fix:python', 'naturalqs:rc::gen2mc', 'mmlu_high_school_chemistry:rc::olmes', 'mmlu_formal_logic:rc::olmes', 'mmlu_world_religions:rc::olmes', 'mmlu_international_law:rc::olmes', 'mmlu_electrical_engineering:rc::olmes', 'mmlu_sociology:rc::olmes', 'mt_mbpp_v2fix:cpp', 'mmlu_elementary_mathematics:rc::olmes', 'mmlu_moral_scenarios:rc::olmes', 'mbpp:3shot::none', 'mmlu_nutrition:rc::olmes', 'lab_bench_dbqa', 'basic_skills:rc::olmes', 'mmlu_high_school_government_and_politics:rc::olmes', 'hellaswag:rc::olmes', 'coqa:rc::gen2mc', 'mt_mbpp_v2fix:haskell', 'mmlu_high_school_us_history:rc::olmes', 'mmlu_econometrics:rc::olmes', 'mmlu_us_foreign_policy:rc::olmes', 'sciq:rc::olmo3', 'medmcqa:rc::none', 'mmlu_prehistory:rc::olmes', 'mmlu_high_school_european_history:rc::olmes', 'mmlu_management:rc::olmes', 'socialiqa:rc::olmes', 'mmlu_computer_security:rc::olmes', 'mt_mbpp_v2fix:swift', 'mmlu_moral_disputes:rc::olmes', 'mt_mbpp_v2fix:rust', 'mmlu_astronomy:rc::olmes', 'minerva_math_precalculus::olmes', 'mmlu_college_medicine:rc::olmes', 'mmlu_anatomy:rc::olmes', 'mt_mbpp_v2fix:typescript', 'codex_humaneval:3shot::none', 'mmlu_medical_genetics:rc::olmes', 'minerva_math_counting_and_probability::o
              ...
              _college_biology:rc::olmes', 'minerva_math_geometry::olmes', 'mmlu_global_facts:rc::olmes', 'mmlu_miscellaneous:rc::olmes', 'basic_skills_string_operations:rc::olmes', 'mt_mbpp_v2fix:php', 'mmlu_high_school_biology:rc::olmes', 'mmlu_human_sexuality:rc::olmes', 'winogrande:rc::olmes', 'mmlu_professional_law:rc::olmes', 'mt_mbpp_v2fix:java', 'mmlu_security_studies:rc::olmes', 'mt_mbpp_v2fix:bash', 'basic_skills_pattern:rc::olmes', 'mmlu_professional_accounting:rc::olmes', 'mt_mbpp_v2fix:ruby', 'mmlu_high_school_macroeconomics:rc::olmes', 'piqa:rc::olmes', 'lambada', 'medqa_en:rc::none', 'mmlu_logical_fallacies:rc::olmes', 'sciriff_yesno:rc::olmes', 'mmlu_high_school_computer_science:rc::olmes', 'basic_skills_arithmetic:rc::olmes', 'mmlu_marketing:rc::olmes', 'mmlu_jurisprudence:rc::olmes', 'arc_challenge:rc::olmes', 'mmlu_high_school_microeconomics:rc::olmes', 'mmlu_high_school_world_history:rc::olmes', 'qasper_yesno:rc::olmes', 'mmlu_public_relations:rc::olmes', 'mt_mbpp_v2fix:matlab', 'csqa:rc::olmes', 'mmlu_high_school_physics:rc::olmes', 'mt_mbpp_v2fix:r', 'squad:rc::gen2mc', 'minerva_math_number_theory::olmes', 'mmlu_business_ethics:rc::olmes', 'mmlu_philosophy:rc::olmes', 'mmlu_virology:rc::olmes', 'minerva_math_algebra::olmes', 'mmlu_conceptual_physics:rc::olmes', 'mmlu_machine_learning:rc::olmes', 'mmlu_college_computer_science:rc::olmes', 'mt_mbpp_v2fix:csharp', 'mt_mbpp_v2fix:scala', 'jeopardy:rc::gen2mc', 'drop:rc::gen2mc', 'minerva_math_intermediate_algebra::olmes'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/allenai/olmix/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv (at revision 33b58d580ba2b439fd9ede85a173188141586db3), [/tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/intermediate_mix_reuse_gap_dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_pdfs/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/theory_validation/weak_mix_reuse_gap_dclm_stackedu/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed0/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed1/ratios.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/metrics.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/metrics.csv), /tmp/hf-datasets-cache/medium/datasets/91025059690133-config-parquet-and-info-allenai-olmix-9328aade/hub/datasets--allenai--olmix/snapshots/33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/ratios.csv (origin=hf://datasets/allenai/olmix@33b58d580ba2b439fd9ede85a173188141586db3/mixture_reuse/real_world/partial_reuse/update1_add_stack_edu_recompute_software_dev_seed2/ratios.csv)]
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

run
string
name
string
index
int64
arc_challenge:rc::olmes
float64
arc_easy:rc::olmes
float64
basic_skills:rc::olmes
float64
basic_skills_arithmetic:rc::olmes
float64
basic_skills_coding:rc::olmes
float64
basic_skills_common_knowledge:rc::olmes
float64
basic_skills_logical_reasoning:rc::olmes
float64
basic_skills_pattern:rc::olmes
float64
basic_skills_string_operations:rc::olmes
float64
codex_humaneval:3shot::none
float64
coqa:rc::gen2mc
float64
csqa:rc::olmes
float64
drop:rc::gen2mc
float64
hellaswag:rc::olmes
float64
jeopardy:rc::gen2mc
float64
lab_bench_dbqa
float64
lab_bench_protocolqa
float64
lambada
float64
mbpp:3shot::none
float64
medmcqa:rc::none
float64
medqa_en:rc::none
float64
minerva_math_algebra::olmes
float64
minerva_math_counting_and_probability::olmes
float64
minerva_math_geometry::olmes
float64
minerva_math_intermediate_algebra::olmes
float64
minerva_math_number_theory::olmes
float64
minerva_math_prealgebra::olmes
float64
minerva_math_precalculus::olmes
float64
mmlu_abstract_algebra:rc::olmes
float64
mmlu_anatomy:rc::olmes
float64
mmlu_astronomy:rc::olmes
float64
mmlu_business_ethics:rc::olmes
float64
mmlu_clinical_knowledge:rc::olmes
float64
mmlu_college_biology:rc::olmes
float64
mmlu_college_chemistry:rc::olmes
float64
mmlu_college_computer_science:rc::olmes
float64
mmlu_college_mathematics:rc::olmes
float64
mmlu_college_medicine:rc::olmes
float64
mmlu_college_physics:rc::olmes
float64
mmlu_computer_security:rc::olmes
float64
mmlu_conceptual_physics:rc::olmes
float64
mmlu_econometrics:rc::olmes
float64
mmlu_electrical_engineering:rc::olmes
float64
mmlu_elementary_mathematics:rc::olmes
float64
mmlu_formal_logic:rc::olmes
float64
mmlu_global_facts:rc::olmes
float64
mmlu_high_school_biology:rc::olmes
float64
mmlu_high_school_chemistry:rc::olmes
float64
mmlu_high_school_computer_science:rc::olmes
float64
mmlu_high_school_european_history:rc::olmes
float64
mmlu_high_school_geography:rc::olmes
float64
mmlu_high_school_government_and_politics:rc::olmes
float64
mmlu_high_school_macroeconomics:rc::olmes
float64
mmlu_high_school_mathematics:rc::olmes
float64
mmlu_high_school_microeconomics:rc::olmes
float64
mmlu_high_school_physics:rc::olmes
float64
mmlu_high_school_psychology:rc::olmes
float64
mmlu_high_school_statistics:rc::olmes
float64
mmlu_high_school_us_history:rc::olmes
float64
mmlu_high_school_world_history:rc::olmes
float64
mmlu_human_aging:rc::olmes
float64
mmlu_human_sexuality:rc::olmes
float64
mmlu_international_law:rc::olmes
float64
mmlu_jurisprudence:rc::olmes
float64
mmlu_logical_fallacies:rc::olmes
float64
mmlu_machine_learning:rc::olmes
float64
mmlu_management:rc::olmes
float64
mmlu_marketing:rc::olmes
float64
mmlu_medical_genetics:rc::olmes
float64
mmlu_miscellaneous:rc::olmes
float64
mmlu_moral_disputes:rc::olmes
float64
mmlu_moral_scenarios:rc::olmes
float64
mmlu_nutrition:rc::olmes
float64
mmlu_philosophy:rc::olmes
float64
mmlu_prehistory:rc::olmes
float64
mmlu_professional_accounting:rc::olmes
float64
mmlu_professional_law:rc::olmes
float64
mmlu_professional_medicine:rc::olmes
float64
mmlu_professional_psychology:rc::olmes
float64
mmlu_public_relations:rc::olmes
float64
mmlu_security_studies:rc::olmes
float64
mmlu_sociology:rc::olmes
float64
mmlu_us_foreign_policy:rc::olmes
float64
mmlu_virology:rc::olmes
float64
mmlu_world_religions:rc::olmes
float64
mt_mbpp_v2fix:bash
float64
mt_mbpp_v2fix:c
float64
mt_mbpp_v2fix:cpp
float64
mt_mbpp_v2fix:csharp
float64
mt_mbpp_v2fix:go
float64
mt_mbpp_v2fix:haskell
float64
mt_mbpp_v2fix:java
float64
mt_mbpp_v2fix:javascript
float64
mt_mbpp_v2fix:matlab
float64
mt_mbpp_v2fix:php
float64
mt_mbpp_v2fix:python
float64
mt_mbpp_v2fix:r
float64
mt_mbpp_v2fix:ruby
float64
mt_mbpp_v2fix:rust
float64
mt_mbpp_v2fix:scala
float64
mt_mbpp_v2fix:swift
float64
mt_mbpp_v2fix:typescript
float64
naturalqs:rc::gen2mc
float64
piqa:rc::olmes
float64
qasper_yesno:rc::olmes
float64
sciq:rc::olmo3
float64
sciriff_yesno:rc::olmes
float64
socialiqa:rc::olmes
float64
squad:rc::gen2mc
float64
winogrande:rc::olmes
float64
md94cxil
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0000
0
1.646324
1.585383
2.525773
2.57052
4.187075
1.758795
0.862001
1.983364
3.792881
2.794646
1.49357
1.835288
2.745005
1.237984
1.973217
5.369895
2.031037
1.298223
3.212887
2.393018
1.905506
1.995711
1.611379
2.155235
2.161568
1.849216
1.759472
1.972276
2.149022
1.797072
1.892648
1.894635
1.827794
1.698308
3.064875
2.499725
3.017362
1.85351
3.052092
2.038457
1.639636
1.625584
2.370182
3.149211
3.234326
2.704817
1.713046
2.743278
2.527913
1.189725
1.661867
1.040064
1.371628
3.489143
1.485567
2.31841
1.595376
1.92378
1.189187
1.205046
1.579559
1.647095
1.079409
1.423674
1.405765
2.316965
1.50715
1.613741
2.269817
2.218462
1.312257
0.495789
1.635006
1.412406
1.858873
1.673932
1.186639
1.761551
1.477064
1.785385
1.134332
1.219606
1.139089
1.622053
2.461649
3.360259
2.659945
2.783422
2.292833
2.979105
3.063964
2.227142
2.909936
2.888345
2.606683
3.212909
3.141748
3.646461
3.318046
3.467647
2.952425
2.832328
2.388889
1.635668
0.411124
1.742954
0.540621
1.621066
1.377888
1.563911
lfagfncq
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0001
1
1.551578
1.47611
2.329994
2.456856
3.644109
1.655133
0.75143
1.787062
3.685375
2.569809
1.460769
1.775269
2.45312
1.195077
1.993773
4.925691
1.804756
1.241782
2.709474
2.209973
1.795644
1.54137
1.335303
1.720943
1.554947
1.463879
1.426544
1.387206
1.706317
1.674985
1.698871
1.863724
1.654871
1.505105
2.798676
2.344501
2.816325
1.712375
2.803572
1.961792
1.585254
1.440991
2.230827
2.970058
2.864573
2.764538
1.498358
2.339761
2.250114
1.244165
1.59946
1.145775
1.429483
3.217008
1.49652
2.163861
1.600994
1.818456
1.282695
1.257117
1.606483
1.653776
1.047861
1.451447
1.4262
1.924232
1.468415
1.645248
2.078692
2.169752
1.354611
0.459857
1.541448
1.448473
1.80096
1.608886
1.165881
1.634812
1.438057
1.77313
1.169075
1.237082
1.214283
1.567222
2.377808
2.809761
2.181146
2.369835
1.982692
2.443879
2.656167
1.909451
2.390662
2.324794
2.293071
2.709408
2.518677
2.887999
2.907986
2.873693
2.539408
2.418884
2.171981
1.520154
0.352079
1.649517
0.504169
1.606982
1.368181
1.533497
y4z0riyy
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0002
2
1.607748
1.578514
2.425506
2.42672
3.661268
1.788236
0.809931
2.048183
3.818695
2.541618
1.543026
1.868237
2.665591
1.274679
2.20962
4.97618
1.77068
1.524921
2.721729
2.38316
1.952692
1.606346
1.402414
1.760682
1.569475
1.530328
1.500124
1.398569
1.735892
1.8177
1.836042
1.879709
1.76433
1.567441
2.735094
2.374938
3.159324
1.752713
2.786797
2.027551
1.613638
1.412528
2.273676
3.183271
2.816553
2.422754
1.593656
2.374214
2.268517
1.358176
1.645751
1.138897
1.470248
3.313185
1.512077
2.188581
1.747042
1.860659
1.330183
1.33084
1.679887
1.775556
1.006464
1.445361
1.487337
1.942798
1.503769
1.632193
2.144809
2.351517
1.381703
0.573335
1.678523
1.502743
1.948237
1.69736
1.039411
1.814656
1.497378
1.843311
1.18491
1.275234
1.279989
1.633495
2.878501
3.208673
2.460486
2.575012
2.215552
2.738374
2.700703
2.213876
2.790653
2.31061
2.729236
2.721629
2.802646
2.925949
3.157487
3.157374
2.797898
2.704848
2.320301
1.69264
0.410349
1.658392
0.469187
1.722518
1.411114
1.604232
ns458nlt
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0003
3
1.592598
1.508796
2.375574
2.423091
3.658671
1.766036
0.736749
2.019787
3.649109
2.575586
1.600635
1.820556
2.82434
1.189352
2.15651
5.002668
1.775845
1.705402
2.705629
2.364445
1.995696
1.559972
1.372107
1.719173
1.549489
1.500042
1.451394
1.385305
1.899177
1.829411
1.800142
1.881672
1.763806
1.585457
2.841394
2.316974
2.86775
1.768765
3.060954
2.04812
1.668258
1.402173
2.262546
2.962074
2.849791
2.412271
1.609791
2.354347
2.291951
1.290036
1.620867
1.172672
1.472461
3.314964
1.492329
2.169616
1.60272
1.887345
1.304461
1.2631
1.655991
1.703847
1.0848
1.573161
1.516384
1.911038
1.512611
1.586173
2.212258
2.278809
1.44319
0.321133
1.677501
1.494398
1.882763
1.743886
1.238128
1.859048
1.517469
1.813228
1.181631
1.278801
1.202967
1.67811
2.522733
3.019432
2.243395
2.380024
2.047218
2.480831
2.6514
1.954169
2.517832
2.275686
2.389455
2.705538
2.541433
2.911503
2.977238
2.887132
2.599479
2.49637
2.284703
1.531774
0.590681
1.65301
0.641731
1.624318
1.387866
1.573483
b4scxlgi
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0004
4
1.577385
1.468055
2.174992
2.414787
2.993944
1.635519
0.749567
1.775089
3.48105
2.05398
1.584825
1.761975
2.679176
1.196012
1.979688
4.916507
1.81002
1.395616
2.197652
2.200266
1.753034
1.51875
1.311889
1.631318
1.527116
1.458974
1.398141
1.352769
2.061706
1.695405
1.754179
1.672469
1.666065
1.514873
2.732841
2.369402
2.931012
1.684312
2.931925
2.007691
1.643592
1.393759
2.233171
3.012364
2.768456
2.430692
1.580427
2.315858
2.326329
1.270632
1.602796
1.191784
1.373051
3.17699
1.40083
2.163977
1.51319
1.840115
1.327438
1.259613
1.564507
1.634632
1.104061
1.470872
1.440322
1.974681
1.420106
1.494285
2.143013
2.131088
1.360185
0.555454
1.559008
1.45506
1.842382
1.669019
1.231845
1.598872
1.415975
1.750413
1.150936
1.207219
1.202178
1.520436
2.370353
2.425377
1.568549
1.665835
1.461188
1.947318
2.24581
1.361861
1.879866
1.958772
1.677968
2.197599
2.177368
2.388334
2.390612
2.304901
1.995699
1.889536
2.224408
1.542341
0.50557
1.635766
0.478963
1.679933
1.388038
1.579555
voddt4rn
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0005
5
1.523034
1.465739
2.294514
2.42176
3.570796
1.532937
0.711573
1.858275
3.671746
2.471125
1.489152
1.748748
2.239349
1.172162
2.028524
4.877175
1.766142
1.329029
2.637573
2.108657
1.630509
1.532837
1.332116
1.711394
1.558474
1.478861
1.423244
1.392533
1.723347
1.578455
1.680736
1.834326
1.583449
1.507453
2.64857
2.342791
2.951506
1.629571
2.67038
2.062931
1.577014
1.439256
2.204981
2.882508
2.788291
2.715514
1.499199
2.244622
2.229534
1.302968
1.640643
1.2239
1.396003
3.098501
1.494114
2.09445
1.510226
1.727504
1.345696
1.299525
1.555066
1.641452
1.152209
1.553184
1.498651
2.038874
1.492076
1.610159
1.957259
2.106203
1.394498
0.335577
1.445897
1.511039
1.864136
1.655604
1.230064
1.52279
1.432125
1.740233
1.18386
1.243144
1.238269
1.493047
2.589904
2.958541
2.070574
2.256485
1.910785
2.393357
2.622684
1.764663
2.36063
2.305171
2.306855
2.637525
2.595036
2.898199
2.886682
2.84827
2.530376
2.339766
2.159277
1.519292
0.433022
1.492759
0.515111
1.654652
1.325346
1.532949
db4fmpwz
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0006
6
1.577354
1.528804
2.364816
2.404054
3.876492
1.613013
0.68993
1.897425
3.707985
2.612138
1.620181
1.82694
2.285907
1.19231
2.063836
4.906208
1.846375
1.585704
2.730053
2.143421
1.664564
1.615425
1.398963
1.742424
1.627475
1.542494
1.476939
1.454746
1.779904
1.588747
1.846232
1.916948
1.618606
1.578512
2.830754
2.301935
2.936083
1.690828
2.719086
2.182628
1.658251
1.374449
2.280134
3.042938
2.957839
2.662811
1.603709
2.427793
2.285654
1.280202
1.6162
1.170549
1.411276
3.130759
1.492747
2.160708
1.484724
1.854931
1.30229
1.257297
1.510837
1.650733
1.071984
1.491549
1.509981
2.030685
1.468721
1.631075
2.12919
2.185023
1.3812
0.326214
1.480876
1.503894
1.917654
1.605691
1.178572
1.545499
1.422981
1.844422
1.148712
1.212185
1.226601
1.512617
2.691577
3.096209
2.31491
2.462251
2.102424
2.599069
2.659053
2.091657
2.696661
2.38107
2.651286
2.730035
2.794056
2.928125
3.098262
3.031537
2.700651
2.599318
2.347308
1.444737
0.405476
1.664265
0.500361
1.698556
1.397387
1.58061
dwm5mawj
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0007
7
1.610461
1.542397
2.109841
2.438322
2.550695
1.635233
0.750353
1.827073
3.457373
1.770595
1.616362
1.811376
2.557887
1.197499
2.250852
4.999956
1.791559
1.3095
1.963495
2.279066
1.842919
1.520698
1.325967
1.617743
1.494464
1.443688
1.403375
1.302448
1.679677
1.704923
1.839033
1.874381
1.722881
1.568659
2.838661
2.242925
2.806491
1.740601
2.939278
1.868107
1.637191
1.362166
2.200207
3.124964
2.790697
2.696361
1.56881
2.329671
2.09845
1.356534
1.712237
1.245913
1.418495
3.179512
1.481629
2.17065
1.645587
1.818232
1.403988
1.367733
1.642218
1.673009
1.20092
1.576926
1.540269
1.901052
1.499354
1.62629
2.17419
2.272646
1.43967
0.391125
1.608603
1.519133
1.977563
1.683355
1.256507
1.76005
1.480783
1.807889
1.178888
1.276525
1.279125
1.612554
2.776621
2.082825
1.375224
1.443688
1.21777
1.756195
2.024973
1.171949
1.649356
1.821746
1.440984
1.963463
1.954064
2.209315
2.158097
2.097513
1.771488
1.684364
2.47028
1.561364
0.437076
1.705123
0.488271
1.674198
1.391035
1.540122
z399ssgq
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0008
8
1.540848
1.473448
2.084667
2.390273
2.528313
1.66663
0.608084
1.84109
3.473614
1.837794
1.524992
1.726051
2.490081
1.192212
1.947389
4.943478
1.751647
1.289626
1.992311
2.222114
1.822803
1.510935
1.302133
1.614738
1.508699
1.434988
1.390129
1.325805
1.666544
1.714961
1.693278
1.964067
1.648504
1.516935
2.59805
2.164658
2.7177
1.672535
2.653175
1.945123
1.568251
1.373773
2.225343
2.920775
2.686912
2.523241
1.555987
2.330738
2.131064
1.222798
1.567993
1.126244
1.362141
2.973493
1.439663
2.01209
1.579354
1.796628
1.244303
1.222018
1.641526
1.679695
1.084516
1.509767
1.451059
1.955523
1.461224
1.571299
2.230109
2.094175
1.385874
0.380955
1.583385
1.455062
1.796865
1.576756
1.201798
1.699706
1.46949
1.76433
1.141034
1.238674
1.164777
1.571515
2.500277
2.20104
1.417864
1.474992
1.283629
1.730637
2.05075
1.193631
1.681836
1.849972
1.461333
1.992238
2.034632
2.22842
2.178291
2.154723
1.747193
1.720348
2.159037
1.551232
0.447031
1.577033
0.431343
1.677191
1.338729
1.543209
4s0c00gh
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0009
9
1.611171
1.51066
2.383875
2.52614
3.750528
1.597003
0.778906
1.969313
3.681361
2.50657
1.567197
1.705783
2.470212
1.192838
1.984527
5.127402
1.875178
1.180846
2.72113
2.248911
1.822902
1.837537
1.53326
1.94651
1.851231
1.716659
1.640999
1.635494
1.733836
1.723363
1.887656
1.883935
1.733976
1.701392
2.860983
2.452964
2.821556
1.81294
2.891823
2.214065
1.609416
1.506614
2.253748
3.004645
3.080392
2.256616
1.682555
2.510013
2.329913
1.259529
1.578099
1.194143
1.461905
3.261426
1.506332
2.272258
1.563186
1.892571
1.315809
1.256801
1.602783
1.633167
1.129591
1.52509
1.429944
2.11104
1.468265
1.622302
2.253682
2.171686
1.34555
0.363397
1.620034
1.434255
1.900807
1.694626
1.25518
1.680009
1.472276
1.78358
1.149446
1.22165
1.222835
1.585815
2.489755
3.099146
2.107026
2.238438
1.919474
2.417622
2.65539
1.758192
2.365869
2.406445
2.14334
2.721061
2.650178
3.024106
2.815244
2.88932
2.48785
2.366992
2.290606
1.51264
0.360091
1.676549
0.416585
1.626175
1.41319
1.533112
8vcq7orj
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0010
10
1.65473
1.598587
2.259711
2.481028
3.175419
1.770845
0.747769
1.779422
3.603785
2.217132
1.481372
1.766493
2.413807
1.187417
2.159358
5.268142
1.995859
1.328862
2.353202
2.559815
2.277506
1.705285
1.443424
1.815613
1.708403
1.59196
1.551219
1.502845
1.87549
2.111112
1.905807
1.867844
1.858972
1.780036
3.014658
2.301712
2.852336
1.82634
3.110433
2.141555
1.68479
1.462636
2.326301
3.137638
2.874559
2.50982
1.762853
2.604302
2.100963
1.288442
1.666585
1.128395
1.419646
3.136043
1.496271
2.275552
1.595317
1.874294
1.281813
1.297191
1.660759
1.667784
1.079646
1.492853
1.507255
2.021039
1.520696
1.642272
2.354716
2.148536
1.394459
0.412057
1.817701
1.496053
1.927307
1.677421
1.169978
2.086971
1.487419
1.781397
1.167962
1.232928
1.191428
1.765792
2.642147
2.631837
1.710619
1.809462
1.689025
2.155943
2.394413
1.526675
2.139872
2.064282
1.953181
2.353125
2.381069
2.516181
2.580011
2.484644
2.200755
2.094179
2.235671
1.531295
0.328244
1.880146
0.414435
1.548926
1.363545
1.539892
gjo4upwr
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0011
11
1.534746
1.4559
2.104541
2.40799
2.505647
1.645091
0.756893
1.853042
3.458584
1.699446
1.527185
1.806076
2.483991
1.179826
1.938057
4.912003
1.771127
1.424297
1.860276
2.240828
1.838087
1.489519
1.30266
1.606676
1.467938
1.419269
1.394033
1.29203
1.630291
1.767714
1.713701
1.887802
1.716867
1.529702
2.70504
2.224476
2.782702
1.680562
2.797178
1.880952
1.5232
1.380726
2.164032
2.976261
2.688878
2.477384
1.524569
2.326397
1.963743
1.208018
1.52057
1.096174
1.377576
3.262042
1.459251
2.077504
1.521192
1.739508
1.240883
1.210404
1.580329
1.682851
1.042631
1.443668
1.4535
1.825782
1.444806
1.625569
2.067979
2.104643
1.35895
0.412533
1.541602
1.463583
1.820148
1.647301
1.162638
1.727179
1.461221
1.750733
1.162657
1.221238
1.164228
1.555579
2.472528
2.094839
1.317944
1.369437
1.2
1.650686
1.941807
1.109411
1.570218
1.74144
1.372325
1.860193
1.925281
2.047638
2.067423
1.993369
1.657454
1.587934
2.177033
1.482394
0.43322
1.602824
0.413186
1.689504
1.338248
1.541166
ywc6ogwe
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0012
12
1.452294
1.325681
2.154096
2.380424
2.988323
1.526595
0.709542
1.790513
3.529181
1.991156
1.501295
1.798446
2.227076
1.220981
2.063332
5.017242
1.754309
1.378599
2.223435
2.309088
1.942373
1.426789
1.257916
1.56357
1.479719
1.356604
1.323849
1.313631
1.694928
1.740805
1.530877
1.822172
1.707787
1.412311
2.65322
2.187905
2.659637
1.634025
2.770512
1.916924
1.53283
1.458298
2.165762
3.039454
2.7162
2.74141
1.44893
2.125575
2.080377
1.288807
1.582769
1.117138
1.325596
3.047024
1.380805
2.020879
1.567458
1.751475
1.298172
1.272434
1.649054
1.672779
1.100942
1.473243
1.462948
1.954225
1.424182
1.49582
2.020992
2.141977
1.336509
0.401604
1.618618
1.474798
1.738586
1.581666
1.164554
1.813251
1.478474
1.71794
1.121209
1.213981
1.174008
1.551644
2.618279
2.512944
1.682634
1.759776
1.497197
2.02254
2.301239
1.397144
1.96205
2.069584
1.710924
2.223378
2.199597
2.512966
2.466017
2.438687
2.068666
1.983924
2.29791
1.578868
0.362659
1.522406
0.415964
1.69979
1.296436
1.559276
zb5cw4pg
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0013
13
1.530098
1.468972
2.305037
2.393951
3.431959
1.717382
0.679334
1.858035
3.74956
2.33839
1.581321
1.737665
2.621519
1.206837
2.077214
4.92695
1.747673
1.421442
2.561594
2.217392
1.735526
1.500679
1.301692
1.644262
1.497553
1.438642
1.392198
1.333205
1.937377
1.624751
1.695968
1.796458
1.68052
1.504028
2.787553
2.295421
2.739707
1.68202
2.786536
1.971613
1.653618
1.393835
2.188736
3.197377
2.775229
2.692909
1.495755
2.286043
2.212859
1.275202
1.612059
1.154178
1.328464
3.066078
1.425193
2.075407
1.480848
1.794699
1.30757
1.288305
1.596324
1.667901
1.091834
1.492039
1.465292
1.927392
1.451027
1.517426
2.121641
2.133502
1.358527
0.391977
1.508334
1.49069
1.872109
1.669615
1.144601
1.644882
1.426854
1.743214
1.17003
1.250521
1.227586
1.574648
2.61874
2.808847
2.012185
2.196514
1.95947
2.394277
2.603454
1.826208
2.320316
2.154908
2.204037
2.561519
2.42336
2.746511
2.857218
2.822041
2.503238
2.345354
2.301988
1.566316
0.44825
1.617219
0.516951
1.606206
1.318117
1.564563
5ygygk6w
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0014
14
1.50556
1.395467
2.204872
2.45741
3.05753
1.566629
0.725257
1.864713
3.557691
2.155134
1.515345
1.791285
2.373321
1.222002
1.937408
5.070065
1.785368
1.54405
2.350225
2.20017
1.78211
1.494532
1.305814
1.62813
1.503691
1.425865
1.392367
1.339791
1.728752
1.662322
1.588626
1.740898
1.624473
1.453542
2.560229
2.230435
2.641186
1.62564
2.67391
1.975086
1.528977
1.428489
2.135686
2.85063
2.759505
2.489231
1.460527
2.228297
2.10252
1.222138
1.578461
1.099365
1.28103
3.03932
1.393536
2.02845
1.524484
1.761952
1.261157
1.236538
1.563945
1.613664
1.051529
1.453597
1.459857
1.940821
1.420674
1.516674
2.035759
2.100891
1.352402
0.539156
1.515977
1.465901
1.793473
1.58751
1.1802
1.640309
1.430729
1.72638
1.144936
1.219128
1.15292
1.517576
2.469672
2.695876
1.719567
1.847112
1.764021
2.143317
2.363013
1.566685
2.12803
2.07339
2.007063
2.350148
2.300209
2.555625
2.658917
2.484177
2.253122
2.116785
2.252093
1.585843
0.439675
1.465577
0.548886
1.695414
1.297853
1.553414
s2kh61fx
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0015
15
1.535784
1.437484
2.276802
2.447181
3.381667
1.711296
0.742284
1.740532
3.637854
2.304251
1.61566
1.82495
2.420379
1.226183
2.049969
4.779287
1.71912
1.410947
2.522703
2.264086
1.827476
1.455608
1.283903
1.602603
1.431937
1.387625
1.369266
1.266245
1.798832
1.690629
1.672128
1.836861
1.699794
1.463676
2.596561
2.23299
2.757105
1.670159
2.580439
1.932323
1.577285
1.360368
2.215833
2.926076
2.570826
2.803773
1.503514
2.188316
2.167934
1.259483
1.698634
1.155022
1.434903
3.045162
1.496116
2.048052
1.515012
1.813198
1.280023
1.244187
1.640273
1.66264
1.093002
1.543606
1.480572
1.825248
1.518066
1.647
2.134139
2.192422
1.396589
0.353163
1.603872
1.511797
1.813492
1.650277
1.23861
1.731805
1.474034
1.811936
1.182887
1.258701
1.202977
1.571704
2.522614
2.790436
1.99475
2.152289
1.943077
2.343657
2.496897
1.818317
2.344629
2.133777
2.285442
2.522659
2.325125
2.738395
2.860246
2.773493
2.508507
2.333219
2.284187
1.595559
0.34849
1.573058
0.471857
1.591313
1.340866
1.568157
thhksf7t
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0016
16
1.557269
1.532598
2.359131
2.456168
3.665631
1.686793
0.761461
1.867055
3.717676
2.579671
1.462395
1.771807
2.318099
1.196324
2.020176
5.269523
1.900262
1.292846
2.772033
2.263078
1.828705
1.642458
1.422861
1.824832
1.67204
1.568848
1.510483
1.482183
2.093437
1.69306
1.750895
1.945861
1.670832
1.585169
2.780564
2.456081
2.887408
1.744959
2.908638
2.116237
1.596124
1.499583
2.220414
2.972028
3.014631
2.374993
1.592673
2.434172
2.246762
1.20834
1.607211
1.091868
1.398741
3.189364
1.471615
2.180305
1.524432
1.84618
1.235577
1.217193
1.574189
1.613018
1.096742
1.464188
1.417453
2.039541
1.52687
1.631991
2.155582
2.137029
1.349383
0.495182
1.524851
1.44176
1.850598
1.658989
1.213718
1.674684
1.465224
1.791098
1.112494
1.222346
1.170985
1.573301
2.306338
3.041892
2.175343
2.299611
1.958466
2.491955
2.731121
1.836764
2.443146
2.447342
2.266635
2.771982
2.622676
2.99721
2.944699
2.923499
2.563779
2.433971
2.116265
1.53088
0.474593
1.695726
0.740171
1.738081
1.350972
1.562537
6u9pb0d5
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0017
17
1.566018
1.48898
2.31173
2.500972
3.521575
1.601381
0.724529
1.828245
3.693677
2.396519
1.477779
1.797684
2.384881
1.190029
1.912567
5.005055
1.795129
1.399966
2.608381
2.153242
1.691531
1.641516
1.415395
1.783182
1.644469
1.560283
1.509436
1.454313
1.916538
1.581832
1.754903
1.889155
1.604146
1.54817
2.695156
2.334496
2.793185
1.678475
2.933628
2.111679
1.610639
1.451246
2.210007
2.957787
2.854647
2.358263
1.560963
2.360114
2.26176
1.208178
1.552658
1.136465
1.322859
3.082534
1.431971
2.136955
1.551571
1.822694
1.204154
1.180257
1.533002
1.626615
1.097522
1.479823
1.521605
2.097648
1.415199
1.544077
2.021289
2.017167
1.368356
0.499395
1.457266
1.474917
1.735095
1.601278
1.233204
1.553512
1.426648
1.736144
1.150288
1.234063
1.184027
1.507784
2.432176
2.969223
2.047273
2.169377
2.00281
2.424594
2.547723
1.842349
2.428984
2.279192
2.301012
2.608293
2.589043
2.756601
2.920024
2.788927
2.458898
2.385146
2.103474
1.52171
0.496183
1.63868
0.741162
1.688056
1.352985
1.565087
sfp9cezr
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0018
18
1.622504
1.547877
1.993183
2.409289
2.009005
1.729234
0.699307
1.729988
3.382278
1.530003
1.577415
1.886021
2.556255
1.246177
1.91707
5.153983
1.832631
1.371534
1.63162
2.444639
2.065638
1.49999
1.301315
1.561629
1.495079
1.412538
1.375727
1.278086
1.76621
1.985899
1.80882
1.877611
1.84364
1.643744
2.894331
2.138232
3.071834
1.826622
2.970483
1.911376
1.64302
1.426398
2.21894
3.085875
2.701612
3.012075
1.644469
2.464445
1.995155
1.178735
1.638362
1.142314
1.447467
3.25734
1.505493
2.187296
1.648331
1.865189
1.208816
1.167364
1.693884
1.756344
1.108091
1.481475
1.523032
1.787032
1.537395
1.655484
2.311388
2.188084
1.367226
0.423159
1.714739
1.481929
1.749767
1.695946
1.212296
1.91531
1.510744
1.778174
1.149524
1.248176
1.184803
1.653802
2.409892
1.732113
1.133272
1.151266
0.975821
1.425534
1.704075
0.909436
1.355107
1.560664
1.172156
1.631598
1.747323
1.832531
1.851241
1.66268
1.41269
1.335657
2.23816
1.608259
0.383364
1.676162
0.57344
1.652328
1.354588
1.585721
qxf73cow
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0019
19
1.587214
1.446169
2.21563
2.434281
3.217438
1.631636
0.726079
1.723965
3.560381
2.229725
1.520886
1.757932
2.256961
1.189054
2.011475
4.962495
1.790644
1.319064
2.377778
2.16374
1.680632
1.577821
1.362851
1.691689
1.556129
1.499043
1.457356
1.382561
1.765096
1.6461
1.797307
1.80795
1.679133
1.531048
2.695405
2.338332
2.883496
1.693714
3.126715
2.045213
1.60998
1.29959
2.276439
3.114241
2.929102
3.071085
1.557065
2.347257
2.247313
1.261621
1.605147
1.112217
1.38964
3.233908
1.480902
2.133285
1.532572
1.912293
1.253486
1.237316
1.561502
1.623747
1.071146
1.46007
1.456693
1.959826
1.421416
1.530573
2.153979
2.194638
1.361651
0.408432
1.523764
1.446825
1.818089
1.668754
1.16659
1.584978
1.431446
1.746687
1.158683
1.212108
1.174321
1.504427
2.574515
2.706197
1.752733
1.910497
1.698582
2.173507
2.442203
1.537289
2.112798
2.096695
1.975831
2.377753
2.375226
2.643965
2.657261
2.532402
2.217332
2.089469
2.216681
1.566445
0.549043
1.655105
0.586038
1.617216
1.344844
1.540641
mz3rdnmu
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0020
20
1.508948
1.42706
2.253428
2.384614
3.255065
1.632716
0.717103
1.891624
3.639448
2.251348
1.591269
1.767829
2.329369
1.213448
2.029591
4.921768
1.703629
1.394562
2.427255
2.198715
1.77398
1.401929
1.252892
1.558155
1.411048
1.355629
1.313367
1.244562
1.642392
1.680162
1.628683
1.853179
1.671021
1.454618
2.549109
2.146094
2.627502
1.625292
2.507762
2.012328
1.552766
1.342365
2.166408
2.958601
2.695179
2.755572
1.467881
2.123413
2.14074
1.312966
1.636596
1.236694
1.412139
2.958021
1.4607
1.957737
1.544311
1.751009
1.351372
1.311208
1.617232
1.683497
1.08902
1.495083
1.54461
1.907605
1.46111
1.604867
2.08994
2.140798
1.429555
0.357445
1.530056
1.525703
1.805118
1.63867
1.165868
1.650976
1.451798
1.752603
1.179566
1.256154
1.275973
1.57439
2.699936
2.796457
1.834065
2.034957
1.709281
2.237755
2.494738
1.602673
2.134956
2.071728
2.024848
2.427239
2.314177
2.675623
2.713151
2.616289
2.306282
2.153066
2.225425
1.54926
0.490832
1.529661
0.641001
1.618857
1.294929
1.566942
iwnlw4oz
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0021
21
1.545926
1.471194
2.334045
2.485806
3.728141
1.604534
0.690939
1.83732
3.657528
2.607495
1.47149
1.722494
2.320217
1.191529
2.009084
4.995108
1.816602
1.350894
2.71707
2.260964
1.850996
1.600988
1.388677
1.755284
1.600736
1.530858
1.477182
1.431112
1.784753
1.735181
1.814745
1.681694
1.643959
1.576388
2.784341
2.339194
2.831786
1.647771
2.716123
2.02348
1.607294
1.34703
2.341515
3.236183
2.977371
2.744314
1.595982
2.388382
2.245104
1.262927
1.60498
1.149967
1.315539
3.187072
1.429886
2.196813
1.521772
1.92587
1.277401
1.264914
1.607114
1.655685
1.078985
1.454363
1.444157
1.959263
1.412593
1.499839
2.176234
2.147934
1.337295
0.499657
1.582013
1.429204
1.86356
1.615104
1.180841
1.701379
1.408094
1.672757
1.143154
1.178888
1.207491
1.60687
2.58773
3.028753
2.380203
2.496836
2.149099
2.60367
2.648751
2.086806
2.622618
2.348861
2.544378
2.716992
2.606115
2.890555
3.096337
2.941216
2.678737
2.565923
2.205474
1.508152
0.405422
1.535585
0.371746
1.532017
1.353382
1.546598
z8xmmtbx
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0022
22
1.558164
1.421205
2.350666
2.426822
3.607641
1.743333
0.781606
1.85757
3.687023
2.496494
1.59345
1.745526
2.417268
1.266771
1.867022
5.01391
1.799141
1.333748
2.65608
2.433825
2.085488
1.513437
1.32624
1.683996
1.508075
1.446028
1.421477
1.347109
1.817673
1.899707
1.656721
1.81642
1.813306
1.49335
2.683531
2.350086
2.866788
1.768572
2.796154
2.097578
1.592947
1.432586
2.190376
2.941687
2.799622
2.406596
1.51896
2.267548
2.176799
1.149244
1.523227
1.062838
1.39189
3.179977
1.476951
2.141928
1.57535
1.813012
1.148017
1.152501
1.68983
1.697356
1.092531
1.433223
1.424654
1.985961
1.491662
1.645836
2.188484
2.090917
1.354145
0.354571
1.72462
1.440539
1.698832
1.657456
1.228273
1.947198
1.519525
1.767916
1.143236
1.200547
1.151312
1.603536
2.484635
3.08976
2.252226
2.43706
2.094465
2.55563
2.677657
1.978816
2.535595
2.347713
2.57091
2.655999
2.542124
2.921178
3.022923
2.946013
2.647158
2.511805
2.238171
1.661414
0.450317
1.569857
0.433283
1.575572
1.285029
1.551086
sbxj6zw7
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0023
23
1.552219
1.488758
2.342923
2.444196
3.732948
1.60303
0.686722
1.928185
3.662457
2.650741
1.581991
1.704432
2.42508
1.208891
1.839953
4.916765
1.753751
1.419226
2.825116
2.377338
1.963301
1.587539
1.378439
1.728505
1.559125
1.498406
1.467482
1.392797
1.918144
1.836813
1.802345
1.836697
1.757939
1.549745
2.842783
2.341391
2.909242
1.712973
3.174786
2.111127
1.664275
1.382483
2.400814
3.16149
2.854197
2.585388
1.585738
2.385083
2.334121
1.139948
1.517884
1.188026
1.424849
3.175264
1.487206
2.215879
1.482132
1.859094
1.134645
1.126955
1.598725
1.760323
1.087825
1.481215
1.502218
2.046645
1.462957
1.597361
2.19358
2.134047
1.38563
0.355779
1.632905
1.479047
1.628785
1.677908
1.24014
1.874747
1.456693
1.749434
1.146493
1.20702
1.183853
1.622462
2.359172
3.318232
2.45247
2.594635
2.21157
2.781184
2.681286
2.146358
2.762823
2.396651
2.726354
2.8251
2.748006
2.991283
3.239759
3.144225
2.870881
2.706754
2.150162
1.600782
0.340026
1.620624
0.412703
1.58521
1.343306
1.542329
2yiyudt2
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0024
24
1.436783
1.332278
2.245645
2.398085
3.419392
1.523601
0.710791
1.797587
3.624412
2.212966
1.551411
1.783873
2.316269
1.219557
1.947947
4.818901
1.712733
1.537431
2.692408
2.330324
1.973296
1.386077
1.210912
1.533824
1.440085
1.318108
1.287646
1.282916
1.906727
1.710672
1.51408
1.815177
1.721513
1.358793
2.557016
2.252579
2.732377
1.623982
2.611299
1.961573
1.524626
1.428241
2.144048
2.919433
2.655864
2.461605
1.376857
2.01333
2.086106
1.220325
1.584175
1.169739
1.435442
3.203363
1.452161
1.989631
1.470915
1.761994
1.238543
1.217238
1.641517
1.657715
1.14126
1.489919
1.388963
1.807661
1.447714
1.642442
2.053327
2.052799
1.3753
0.56757
1.588208
1.509337
1.673719
1.687229
1.272942
1.882198
1.463116
1.717726
1.165353
1.222643
1.182502
1.552756
2.537071
3.125751
2.161371
2.349378
2.078622
2.507072
2.70907
1.952466
2.498543
2.34804
2.465304
2.692344
2.590532
3.049682
2.960473
2.956638
2.555205
2.44078
2.16381
1.584711
0.408562
1.50335
0.491957
1.664927
1.315756
1.581408
qdzeuhxg
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0025
25
1.616085
1.523843
2.2833
2.416017
3.387127
1.649807
0.703851
1.806757
3.736243
2.286771
1.500183
1.82839
2.342234
1.224779
2.151802
5.310151
1.87081
1.689161
2.42189
2.273763
1.883089
1.714116
1.455119
1.868668
1.759437
1.640916
1.561707
1.590402
1.666417
1.786028
1.886707
1.812222
1.664553
1.628936
2.80388
2.29656
2.899572
1.75413
2.951858
2.090482
1.605154
1.34997
2.227961
2.91295
2.916077
2.30036
1.67231
2.450063
2.251578
1.283675
1.629846
1.140056
1.257798
3.2587
1.311654
2.219352
1.655858
1.821165
1.305371
1.274507
1.628741
1.746664
1.032395
1.4663
1.539711
2.095306
1.381587
1.464928
2.18358
2.194968
1.393546
0.374253
1.638733
1.474573
1.891627
1.600576
1.162186
1.727175
1.437822
1.687708
1.148342
1.231184
1.192593
1.595253
2.697067
2.756583
1.741537
1.905078
1.6949
2.188978
2.439289
1.519599
2.139131
2.124097
1.903841
2.421799
2.415373
2.629288
2.670477
2.56169
2.209307
2.138423
2.254217
1.538917
0.451689
1.686291
0.517973
1.705981
1.383424
1.561724
t7j16c2a
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0026
26
1.556446
1.480165
2.316194
2.450684
3.794759
1.62284
0.726325
1.778095
3.524464
2.573464
1.493642
1.798924
2.318186
1.197776
1.991949
5.148626
1.875211
1.406026
2.822713
2.281574
1.82613
1.609068
1.388818
1.780416
1.618539
1.521159
1.479822
1.441284
1.852659
1.719597
1.767221
1.843793
1.720377
1.560495
2.926931
2.340974
2.907063
1.749443
2.99717
1.937002
1.652263
1.44987
2.316908
3.296933
2.980285
3.0421
1.583587
2.4402
2.200234
1.206401
1.615638
1.093899
1.399232
3.100532
1.469642
2.210411
1.498593
1.85358
1.235352
1.233761
1.592707
1.670032
1.073879
1.469883
1.417616
2.017152
1.468167
1.599512
2.279118
2.152022
1.313235
0.588692
1.568954
1.467858
1.85633
1.666046
1.164851
1.719616
1.432544
1.752783
1.162353
1.194189
1.181992
1.588493
2.289541
2.92064
2.242547
2.418356
2.036027
2.515916
2.667587
1.914212
2.473116
2.447032
2.328804
2.822686
2.601768
3.052679
2.95846
2.94494
2.589386
2.463648
2.29771
1.577087
0.394125
1.60926
0.559318
1.586893
1.344942
1.552384
g0vvgj8d
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0027
27
1.669746
1.635777
2.289556
2.483412
3.290536
1.716063
0.732573
1.919296
3.595456
2.176944
1.460827
1.974872
2.450351
1.200683
2.187681
5.390016
2.050315
1.470276
2.408206
2.401624
1.976485
1.90791
1.54653
1.989839
1.992653
1.760022
1.675308
1.775188
2.040743
1.901223
1.938905
1.870619
1.768856
1.751674
3.198425
2.370209
2.870677
1.850318
2.878039
2.086258
1.70235
1.509271
2.401829
3.203061
3.180757
2.201702
1.746066
2.852927
2.218435
1.312062
1.75976
1.081603
1.405435
3.322301
1.485698
2.281443
1.64235
1.824734
1.257667
1.284574
1.653144
1.762513
0.994637
1.403578
1.499739
2.209305
1.509881
1.65199
2.320413
2.273024
1.361105
0.514088
1.654155
1.501368
1.982513
1.640507
1.019358
1.815891
1.4711
1.783787
1.175971
1.228258
1.237905
1.671315
2.706087
2.602219
1.691606
1.78013
1.525963
2.036668
2.318863
1.414046
2.003912
2.199867
1.763056
2.408191
2.343977
2.696752
2.5235
2.522126
2.084551
2.005698
2.30253
1.502768
0.439466
1.884425
0.630659
1.780535
1.426347
1.572379
rr4nsvdx
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0028
28
1.511252
1.410073
2.282943
2.440852
3.585378
1.540357
0.709283
1.774908
3.646882
2.359729
1.523403
1.805829
2.537998
1.170288
2.011809
5.058787
1.779565
1.263061
2.533245
2.281174
1.877511
1.465618
1.305327
1.65803
1.488617
1.417638
1.378046
1.333608
1.978946
1.760659
1.669122
1.892117
1.682759
1.484017
2.664091
2.191929
2.750347
1.65926
2.72537
2.126119
1.531007
1.42236
2.169672
2.996809
2.684285
2.731371
1.496148
2.245111
2.172664
1.248819
1.631129
1.182946
1.328394
3.070591
1.432591
2.064129
1.520036
1.754737
1.326439
1.264843
1.603479
1.639943
1.132205
1.497775
1.407378
1.986401
1.466728
1.579453
2.109719
2.192878
1.348624
0.509017
1.602102
1.431467
1.824284
1.634891
1.202805
1.752313
1.473651
1.754648
1.181903
1.237901
1.238038
1.584797
2.305624
2.944039
2.0006
2.16644
1.890389
2.345426
2.571523
1.756314
2.378983
2.252679
2.353011
2.533163
2.472988
2.82344
2.852345
2.695594
2.514278
2.385352
2.224597
1.510496
0.393891
1.54695
0.441871
1.674818
1.362202
1.541751
t1v65ree
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0029
29
1.628465
1.581513
2.470037
2.40488
4.202759
1.743316
0.76529
1.947716
3.756259
2.934528
1.488096
1.835524
2.40416
1.235943
2.086663
5.212576
1.886589
1.625213
3.134875
2.11504
1.595585
1.73869
1.498241
1.938884
1.767026
1.639998
1.602122
1.583163
1.872952
1.496755
1.841239
1.812725
1.590874
1.597173
2.827392
2.49319
2.873191
1.71492
2.785107
2.295852
1.684101
1.423204
2.355021
3.010221
3.323664
2.353507
1.605523
2.474823
2.354472
1.211682
1.585284
1.066385
1.386127
3.257509
1.470609
2.25662
1.550392
1.922526
1.195551
1.220672
1.527902
1.601054
1.069619
1.480027
1.438829
2.148484
1.509425
1.668215
2.029353
2.151004
1.354657
0.366143
1.460736
1.513648
1.91587
1.692762
1.203343
1.482107
1.429631
1.837989
1.108804
1.197201
1.166283
1.495644
2.525387
3.509985
2.755067
2.852927
2.441733
2.925637
2.994617
2.37363
2.985971
2.687201
2.916332
3.134828
3.019941
3.322196
3.443295
3.385882
3.020847
2.879577
2.30906
1.565229
0.33278
1.651368
0.508106
1.693054
1.383562
1.583943
zvc0nyut
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0030
30
1.705219
1.654891
2.5182
2.635579
4.124085
1.716504
0.799063
1.959314
3.874656
2.978658
1.584799
1.838871
2.38494
1.266945
1.968595
5.700928
2.385225
1.148374
3.265139
2.882074
2.718146
2.751712
2.042764
2.720354
2.890014
2.404818
2.313576
2.750105
1.9604
2.289357
1.903706
1.969376
2.168709
2.06813
3.470697
2.65581
3.068007
2.233628
3.40305
2.466325
1.762981
1.643193
2.702093
3.113788
3.033244
2.785948
1.913258
3.102912
2.593043
1.252774
1.738305
1.289237
1.567784
3.652573
1.648342
2.507001
1.71327
2.193987
1.335113
1.266769
1.818553
1.872493
1.221647
1.447468
1.360736
2.341788
1.593397
1.796192
2.668284
2.295468
1.371384
0.600633
2.229194
1.416196
1.918601
1.78867
1.270019
2.43796
1.586517
1.867796
1.197926
1.257933
1.266689
1.960647
2.528406
3.72707
2.957941
3.14598
2.493258
3.262966
3.135283
2.509788
3.182859
2.959508
3.083034
3.265009
3.428546
3.663903
3.577336
3.634588
3.172588
3.065105
2.32859
1.643178
0.550096
1.860271
0.920227
1.648596
1.445408
1.52507
0fyi7huk
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0031
31
1.643498
1.592868
2.418998
2.443412
3.888811
1.781226
0.831512
1.878126
3.690903
2.77382
1.561027
1.845639
2.197551
1.237184
2.04046
5.099838
1.957853
1.268206
2.899344
2.489683
2.062053
1.835197
1.513504
1.983021
1.878846
1.712385
1.644556
1.670893
2.175054
1.962354
1.880078
1.823365
1.864776
1.773329
2.942462
2.531863
2.856767
1.910304
3.09278
2.195456
1.703307
1.535
2.395415
3.063075
3.015383
2.55185
1.750517
2.66495
2.331023
1.251909
1.650656
1.09103
1.371905
3.254066
1.467389
2.272082
1.639333
1.886611
1.219677
1.227781
1.673019
1.76179
1.089897
1.466924
1.421114
2.19647
1.451032
1.581624
2.412783
2.141527
1.363867
0.653443
1.78331
1.473202
1.920986
1.662697
1.178731
1.924316
1.493339
1.783165
1.138739
1.239697
1.167705
1.661731
2.624327
3.293915
2.4889
2.698751
2.237511
2.662503
2.803123
2.121666
2.647438
2.584321
2.606158
2.899279
2.812687
3.14149
3.102578
3.104664
2.70037
2.615251
2.247456
1.651476
0.507717
1.773255
0.50515
1.587607
1.337056
1.5695
szxf7cy3
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0032
32
1.580224
1.522765
2.269828
2.39379
3.284296
1.648303
0.742998
1.766814
3.782766
2.253197
1.497812
1.739863
2.467734
1.193851
1.948975
5.006757
1.802804
1.304202
2.442766
2.402256
1.992613
1.581402
1.362331
1.711149
1.573161
1.514258
1.454022
1.391556
1.87305
1.846696
1.748198
1.984512
1.726511
1.603362
2.759993
2.323453
3.056955
1.747761
2.709334
1.904613
1.604071
1.427965
2.284421
3.212616
2.886447
2.340807
1.599966
2.43246
2.266295
1.19157
1.60079
1.104684
1.396012
3.280609
1.474344
2.158258
1.571151
1.912303
1.210513
1.207297
1.646883
1.733973
1.067257
1.470829
1.426782
2.021869
1.512536
1.62587
2.29576
2.108921
1.358633
0.297481
1.640981
1.460576
1.769163
1.719425
1.195527
1.846861
1.483879
1.760623
1.145829
1.250063
1.167627
1.646731
2.412711
2.629834
1.789875
1.933413
1.678891
2.142287
2.441231
1.533514
2.142884
2.122051
1.970357
2.442697
2.3953
2.61677
2.612257
2.510249
2.231828
2.14779
2.080856
1.486205
0.363129
1.674262
0.383805
1.653055
1.286659
1.569905
klzcbs2i
backfill-5xC-30m-dclm-s2pdf-flat-83ea88e1-0033
33
1.485282
1.38181
2.314719
2.391427
3.625783
1.570455
0.717613
1.808916
3.774117
2.443767
1.581562
1.731599
2.323977
1.197298
1.904533
4.82688
1.760408
1.309155
2.593699
2.223002
1.812926
1.438099
1.265957
1.586657
1.454095
1.372706
1.338281
1.283732
1.911134
1.652797
1.629777
1.835066
1.668412
1.431716
2.668462
2.239924
2.62272
1.646895
2.636435
2.006904
1.532002
1.388643
2.095295
2.933455
2.697373
2.21545
1.431117
2.134553
2.215735
1.268602
1.567797
1.183771
1.425133
3.032887
1.48725
2.051762
1.518932
1.71638
1.312801
1.265285
1.54695
1.618629
1.076134
1.464189
1.424285
1.929633
1.4641
1.560571
2.061401
2.149382
1.351989
0.382257
1.536764
1.443288
1.761961
1.637555
1.17925
1.66056
1.419313
1.728809
1.124191
1.217507
1.227214
1.550259
2.518262
2.981843
2.133599
2.328906
1.996492
2.409778
2.586601
1.876989
2.385114
2.256022
2.342808
2.593663
2.532789
2.859927
2.917742
2.842941
2.498158
2.376819
2.240696
1.526904
0.445621
1.539642
0.471138
1.603825
1.344939
1.533549
End of preview.

Olmix Swarm Datasets

This repository contains proxy run swarm datasets for data mixing using Olmix. Each swarm consists of multiple proxy training runs with different domain mixture ratios and their corresponding evaluation metrics across various downstream tasks. The swarm datasets here are based on 30M parameter proxy models trained on 3B tokens.

For more details, see our paper: Olmix: A Framework for Data Mixing Throughout LM Development

Swarms

The dataset contains 36 swarms. Each swarm consists of ratios.csv, metrics.csv, and meta.json.

dclm_swarm/                          # Main DCLM swarm (1 swarm)
β”œβ”€β”€ study/
β”‚   β”œβ”€β”€ rq2/                         # Swarm size experiments (3 swarms)
β”‚   β”‚   β”œβ”€β”€ dclm_swarm_18_topics/
β”‚   β”‚   β”œβ”€β”€ dclm_swarm_12_topics/
β”‚   β”‚   └── dclm_swarm_6_topics/
β”‚   β”œβ”€β”€ rq3/                         # Swarm distribution experiments (5 swarms)
β”‚   β”‚   β”œβ”€β”€ dclm_dense_swarm/
β”‚   β”‚   β”œβ”€β”€ source_level_sparse_swarm/
β”‚   β”‚   β”œβ”€β”€ source_level_dense_swarm/
β”‚   β”‚   β”œβ”€β”€ dclm_swarm_strong_prior/
β”‚   β”‚   └── dclm_swarm_weak_prior/
β”‚   └── rq6/                         # Constrained swarm for enforcing repetition (1 swarm)
β”‚       └── dclm_constrained_swarm/
└── mixture_reuse/
    β”œβ”€β”€ real_world/
    β”‚   β”œβ”€β”€ full_reuse/              # Full mixture reuse (12 swarms)
    β”‚   β”‚   β”œβ”€β”€ update1_add_stack_edu_seed{0,1,2}/
    β”‚   β”‚   β”œβ”€β”€ update2_add_more_sources_seed{0,1,2}/
    β”‚   β”‚   β”œβ”€β”€ update3_revise_pdfs_seed{0,1,2}/
    β”‚   β”‚   └── update5_partition_pdfs_seed{0,1,2}/
    β”‚   β”œβ”€β”€ partial_reuse/           # Partial mixture reuse (6 swarms)
    β”‚   β”‚   β”œβ”€β”€ update1_add_stack_edu_recompute_software_dev_seed{0,1,2}/
    β”‚   β”‚   └── update5_partition_pdfs_recompute_mix_seed{0,1,2}/
    β”‚   β”œβ”€β”€ swarm_reuse/             # Swarm reuse baseline (1 swarm)
    β”‚   └── full_recomputation/      # Full recomputation baseline (1 swarm)
    └── theory_validation/           # Theoretical validation (6 swarms)
        β”œβ”€β”€ dclm_stackedu/
        β”œβ”€β”€ dclm_pdfs/
        β”œβ”€β”€ weak_mix_reuse_gap_dclm_stackedu/
        β”œβ”€β”€ intermediate_mix_reuse_gap_dclm_stackedu/
        β”œβ”€β”€ weak_mix_reuse_gap_dclm_pdfs/
        └── intermediate_mix_reuse_gap_dclm_pdfs/
  • dclm_swarm/ (1 swarm): 128 proxy runs over 24 DCLM topics (sparse); the primary swarm used throughout the paper.

  • study/ (9 swarms):

    • RQ2 β€” Swarm Size (3 swarms, Section 3.2): How many proxy runs are needed for m domains? Swarms over subsets of DCLM topics: 18 topics (129 runs total), 12 topics (130 runs total), and 6 topics (125 runs total).
    • RQ3 β€” Swarm Distribution (5 swarms, Section 3.3): Comparing swarm configurations over 24 DCLM topics or 7 sources, varying density (sparse vs. dense) and Dirichlet prior strength (weak vs. strong).
    • RQ6 β€” Constrained Optimization (1 swarm, Section 3.6): 128 proxy runs satisfying repetition constraints for R=6T requested tokens and k=4 repetition factor.
  • mixture_reuse/real_world/ (20 swarms, Section 5.1, Figure 10):

    • Full reuse (12 swarms): The entire previous mixture is reused as a single virtual domain. 3 seeds each for 4 updates:
      • Update 1 (add Stack-Edu): swarms over DCLM (1 virtual domain) and Stack-Edu
      • Update 2 (add more sources): swarms over DCLM+Stack-Edu (1 virtual domain) and new sources
      • Update 3 (revise PDFs): swarms over revised PDF source and unchanged domains (1 virtual domain)
      • Update 5 (partition PDFs): swarms over partitioned PDF topics and unchanged domains (1 virtual domain)
    • Partial reuse (6 swarms): Only the unaffected portion of the mixture is reused; affected domains are recomputed. 3 seeds each for 2 updates:
      • Update 1 (add Stack-Edu, recompute software development): swarms over DCLM \ software development (1 virtual domain), software development, and Stack-Edu
      • Update 5 (partition PDFs, recompute mix): swarms over PDF topics and sources
    • Swarm reuse (1 swarm): Proxy runs accumulated from all swarms compatible with the final 64-domain set.
    • Full recomputation (1 swarm): 512 proxy runs over the final 64 domains after all 5 updates. Subsampled at various budgets to measure full recomputation's performance as a function of run count.
  • mixture_reuse/theory_validation/ (6 swarms, Section 5.2): Experiments validating mixture reuse theory.

    • DCLM+StackEdu and DCLM+PDFs: Swarms for the full recomputation baseline in two settings.
    • Weak/Intermediate mix reuse gap (DCLM+StackEdu): Swarms over DCLM (reusing a weak or intermediate mix) and Stack-Edu.
    • Weak/Intermediate mix reuse gap (DCLM+PDFs): Swarms over DCLM (reusing a weak or intermediate mix) and PDFs.

Data Format

ratios.csv

  • run, name, index: run identifier, full run name, and sequential index
  • One column per domain: mixture weight (proportion of training tokens) assigned to that domain. Domain names follow the format source:topic (e.g., dclm:politics, s2pdf:science_tech, stack-edu:Python), where the prefix is the data source and the suffix is the topic within that source. Single-part names (e.g., wikipedia, arxiv) are single-domain sources.

metrics.csv

  • run, name, index: same identifiers as in ratios.csv
  • One column per evaluation task: the model's bits-per-byte (BPB) score on that task. Lower BPB is better. Tasks span a wide range of benchmarks including MMLU, ARC, coding (HumanEval, MBPP), math (Minerva), and reading comprehension.

meta.json

Each meta.json includes swarm_id, description, category, notes, files, and two fields used when configuring a fit in the Olmix repository:

  • relative_sizes: The natural distribution of the domains, used in the KL penalty.
  • token_counts: The final token count for each domain when constructing the repetition constraint.

Usage

The primary use of these datasets is as input to the Olmix GitHub repository for constructing a proposed mix. Download ratios.csv, metrics.csv, and meta.json from the desired swarm folder and point to them in your Olmix config .yaml file (see example config).

The dataset is also accessible via the HuggingFace datasets library (load_dataset("allenai/olmix")) for exploration purposes.

Resources

Citation

If you use these datasets in your research, please cite:

@article{chen2026olmix,
  title={Olmix: A Framework for Data Mixing Throughout LM Development},
  author={Chen, Mayee F and Murray, Tyler and Heineman, David and Jordan, Matt and Hajishirzi, Hannaneh and Re, Christopher and Soldaini, Luca and Lo, Kyle},
  journal={arXiv preprint arXiv:2602.12237},
  year={2026}
}

License

This dataset is released under the Apache 2.0 License.

Downloads last month
238

Papers for allenai/olmix