Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall
尚未评估计划受阻指标结果claim_t3_teacher_top1
Mid-training ablation results for Teacher Top-1 Labels relative difference across Reasoning, Factual Recall, and Knowledge task groups.
来源:source_paper:Table 3, Page 9
报告指标与观测值
delta_percentage_points
t3_teacher_top1_reasoning
论文报告 -6.4 percentage_points
实际观测 — percentage_points
delta_percentage_points
t3_teacher_top1_factual_recall
论文报告 1.3 percentage_points
实际观测 — percentage_points
delta_percentage_points
t3_teacher_top1_knowledge
论文报告 -2.8 percentage_points
实际观测 — percentage_points
Assessment(0)
这条 Claim 暂无已发布的不可变 Assessment。
关联运行(0)
这条 Claim 暂无关联执行。