open_dutch_llm_leaderboard / evals /truthfulqa /truthfulqa_nl_neural-chat-7b-v3-1.json
Bram Vanroy
add evaluations
9c8bb6d
raw
history blame
564 Bytes
{
"results": {
"truthfulqa_nl": {
"mc1": 0.3668789808917197,
"mc1_stderr": 0.017212610850337793,
"mc2": 0.5149795684141129,
"mc2_stderr": 0.016374624397865214
}
},
"versions": {
"truthfulqa_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=Intel/neural-chat-7b-v3-1,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": {}
}
}