lewtun's picture
lewtun HF staff
Upload eval_results/mistralai/Mixtral-8x7B-Instruct-v0.1/main/arc/results_2024-03-02T18-33-34.606516.json with huggingface_hub
3df4c58 verified
raw
history blame
2.33 kB
{
"config_general": {
"lighteval_sha": "?",
"num_fewshot_seeds": 1,
"override_batch_size": 1,
"max_samples": null,
"job_id": "",
"start_time": 2357016.821755239,
"end_time": 2357428.775015857,
"total_evaluation_time_secondes": "411.953260618262",
"model_name": "mistralai/Mixtral-8x7B-Instruct-v0.1",
"model_sha": "1e637f2d7cb0a9d6fb1922f305cb784995190a83",
"model_dtype": "torch.bfloat16",
"model_size": "87.49 GB",
"config": null
},
"results": {
"lighteval|arc:challenge|25": {
"acc": 0.6979522184300341,
"acc_stderr": 0.013417519144716417,
"acc_norm": 0.7201365187713311,
"acc_norm_stderr": 0.013119040897725922
}
},
"versions": {
"lighteval|arc:challenge|25": 0
},
"config_tasks": {
"lighteval|arc:challenge": {
"name": "arc:challenge",
"prompt_function": "arc",
"hf_repo": "ai2_arc",
"hf_subset": "ARC-Challenge",
"metric": [
"loglikelihood_acc",
"loglikelihood_acc_norm_nospace"
],
"hf_avail_splits": [
"train",
"test"
],
"evaluation_splits": [
"test"
],
"few_shots_split": null,
"few_shots_select": "random_sampling_from_train",
"generation_size": 1,
"stop_sequence": [
"\n"
],
"output_regex": null,
"frozen": false,
"suite": [
"lighteval",
"arc"
],
"original_num_docs": 1172,
"effective_num_docs": 1172
}
},
"summary_tasks": {
"lighteval|arc:challenge|25": {
"hashes": {
"hash_examples": "17b0cae357c0259e",
"hash_full_prompts": "d44b17600302e2b8",
"hash_input_tokens": "0149ae15d3f432fa",
"hash_cont_tokens": "289aa98c400841d8"
},
"truncated": 0,
"non_truncated": 1172,
"padded": 4660,
"non_padded": 27,
"effective_few_shots": 25.0,
"num_truncated_few_shots": 0
}
},
"summary_general": {
"hashes": {
"hash_examples": "aaa6929c6d3771fb",
"hash_full_prompts": "b29ea0f4ca87b218",
"hash_input_tokens": "6064dbfd3e9d9140",
"hash_cont_tokens": "a2928b768b852bf8"
},
"truncated": 0,
"non_truncated": 1172,
"padded": 4660,
"non_padded": 27,
"num_truncated_few_shots": 0
}
}