arXivBoyan Li, Bingsen Chen, Chenghao Yang, Ping Nie, Chen Zhao, Xi YeThu, Sep 3, 2026, 10:14 AM PDT
score 16.9
Two-step training beats mixing tricks for AI reasoning models
Original: Sequential Beats Joint: On the Interplay between On-Policy Distillation and RLVR
Source: arxiv.org ↗
Writing ELI5 summary…