open_dutch_llm_leaderboard / evals /truthfulqa /truthfulqa_nl_gpt-neo-125M-dutch.json
Bram Vanroy
add evaluations
90df5ea
raw
history blame
572 Bytes
{
"results": {
"truthfulqa_nl": {
"mc1": 0.25605095541401274,
"mc1_stderr": 0.015587510684642944,
"mc2": 0.4305171468716022,
"mc2_stderr": 0.015454035132644787
}
},
"versions": {
"truthfulqa_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=yhavinga/gpt-neo-125M-dutch,use_accelerate=True,device_map_option=auto,dtype=bfloat16,load_in_8bit=True",
"batch_size": "auto",
"device": "cuda",
"no_cache": false,
"limit": null,
"bootstrap_iters": 100000,
"description_dict": {}
}
}