{ "protocol": "lm-eval-harness 0.4.13, base-model protocol, HumanEval fence-stripped", "cut": "25+15", "params_b": 12.57, "gflop_per_token": 24.37, "compute_pct_of_larger_parent": 85, "smaller_parent": { "name": "Qwen3-8B-Base", "humaneval": 0.677, "gsm8k_cot": 0.856, "mmlu": 0.7683, "bbh_cot_fewshot": 0.7742, "hellaswag": 0.7859, "arc_easy": 0.7984, "arc_challenge": 0.5717, "openbookqa": 0.416, "winogrande": 0.723, "piqa": 0.7943, "lambada_openai": 0.7105, "mbpp": 0.686, "boolq": 0.8321 }, "model": { "humaneval": 0.671, "gsm8k_cot": 0.8802, "mmlu": 0.7888, "bbh_cot_fewshot": 0.7576, "hellaswag": 0.791, "arc_easy": 0.7765, "arc_challenge": 0.5606, "openbookqa": 0.458, "winogrande": 0.7632, "piqa": 0.8025, "lambada_openai": 0.7068, "mbpp": 0.712, "boolq": 0.8722 }, "larger_parent": { "name": "Qwen3-14B-Base", "humaneval": 0.72, "gsm8k_cot": 0.8704, "mmlu": 0.8053, "bbh_cot_fewshot": 0.7877, "hellaswag": 0.8136, "arc_easy": 0.8194, "arc_challenge": 0.5879, "openbookqa": 0.48, "winogrande": 0.7466, "piqa": 0.8243, "lambada_openai": 0.7419, "mbpp": 0.748, "boolq": 0.8651 } }