open_dutch_llm_leaderboard / evals /truthfulqa /truthfulqa_nl_Orca-2-13b.json
Bram Vanroy
add evaluations
4a78e88
raw
history blame
560 Bytes
{
"results": {
"truthfulqa_nl": {
"mc1": 0.34394904458598724,
"mc1_stderr": 0.016965154087337654,
"mc2": 0.4961166743388786,
"mc2_stderr": 0.016446183810227374
}
},
"versions": {
"truthfulqa_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=microsoft/Orca-2-13b,use_accelerate=True,device_map_option=auto,dtype=bfloat16,load_in_8bit=True",
"batch_size": 2,
"device": "cuda",
"no_cache": false,
"limit": null,
"bootstrap_iters": 100000,
"description_dict": {}
}
}