Offline reinforcement learning with verifiable rewards (RLVR) trains on a fixed pool of sampled responses and typically supervises every token of every response, although responses and tokens are not equally informative. We study which ones to supervise through the Gibbs-optimal policy of KL-regularized RLVR, which off...
Xu-Liang Huang, Tian-Xiang Zhao, Cheng-Wei Qin et al.· 0 citations
Non-prefix KV cache reuse is a promising approach for improving inference efficiency in multi-agent systems as agents share the same context blocks. However, when agents have conflicting tasks, directly reusing non-prefix KV caches across agents can severely degrade model performance, causing up to a 90% drop. Existing...
Zhi-Wei Chen, Zhong-Tao Rao, Wai-Kit Xiu et al.· 0 citations
On-policy distillation (OPD) improves large language model reasoning by training students on their own rollouts with dense token-wise supervision from the teacher. However, token-wise OPD does not explicitly provide a coherent alternative reasoning step showing how the student's step could be revised to improve subsequ...
Yu-Xiang Zhang, Ding Cao, Shu-Ting Cui et al.· 0 citations
DivOPD, a simple learner-side batch-selection method that spreads a fixed turn budget across more rollouts and, within each rollout, prioritizes turns with larger cumulative teacher-student disagreement, is introduced.
Han-Yang Wang, Ze-Yuan Liu, Zheng-Yu Chen et al.· 0 citations
SAGE is a training-free inference-time framework that structures LLM strategic reasoning around three coordinated operations: anchor, adapt, and recalibrate, and distills strategically related interactions into counterfactual hypotheses about previously missing considerations, allowing past experience to recalibrate th...