open_dutch_llm_leaderboard / evals /truthfulqa /truthfulqa_nl_gpt-neo-1.3B-dutch.json
Bram Vanroy
add evaluations
90df5ea
raw
history blame
573 Bytes
{
"results": {
"truthfulqa_nl": {
"mc1": 0.26496815286624203,
"mc1_stderr": 0.015761294574831355,
"mc2": 0.42842315532854475,
"mc2_stderr": 0.015376698164939607
}
},
"versions": {
"truthfulqa_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=yhavinga/gpt-neo-1.3B-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": {}
}
}