评分 6 · 来源：cs.LG updates on arXiv.org · 发布于 2026-04-30
评分依据：Combines privileged on-policy info with adaptive interpolation for reasoner training, addresses credit assignment challenge
PAINT: Partial-Solution Adaptive Interpolated Training for Self-Distilled Reasoners
发布
采集
学术前沿 6.0 分
— Combines privileged on-policy info with adaptive interpolation for reasoner training, addresses credit assignment challenge