{ "protocol": "lm-eval-harness 0.4.13, base-model protocol, HumanEval fence-stripped", "cut": "18+10", "params_b": 1.53, "gflop_per_token": 2.68, "compute_pct_of_larger_parent": 73, "smaller_parent": { "name": "Qwen3-0.6B-Base", "humaneval": 0.287, "gsm8k_cot": 0.5951, "mmlu": 0.5271, "bbh_cot_fewshot": 0.4061, "hellaswag": 0.5384, "arc_easy": 0.5791, "arc_challenge": 0.3848, "openbookqa": 0.344, "winogrande": 0.5872, "piqa": 0.7013, "lambada_openai": 0.5372, "mbpp": 0.39, "boolq": 0.6972 }, "model": { "humaneval": 0.409, "gsm8k_cot": 0.6914, "mmlu": 0.6056, "bbh_cot_fewshot": 0.4732, "hellaswag": 0.5951, "arc_easy": 0.6604, "arc_challenge": 0.407, "openbookqa": 0.374, "winogrande": 0.6354, "piqa": 0.7258, "lambada_openai": 0.5742, "mbpp": 0.456, "boolq": 0.7771 }, "larger_parent": { "name": "Qwen3-1.7B-Base", "humaneval": 0.494, "gsm8k_cot": 0.7293, "mmlu": 0.6249, "bbh_cot_fewshot": 0.529, "hellaswag": 0.6648, "arc_easy": 0.6864, "arc_challenge": 0.4462, "openbookqa": 0.388, "winogrande": 0.644, "piqa": 0.7579, "lambada_openai": 0.633, "mbpp": 0.566, "boolq": 0.7902 } }