It is revealed that off-the-shelf LRMs exhibit a confidence prior heavily concentrated on a few high values, which persists throughout RL, and it is proved that this concentration suppresses policy gradient updates for rarely sampled confidence values and inflates the lower bound on expected Brier risk.
Shuo-Yuan Wang, Bei-Er Luo, Hao Zeng et al.· 0 citations
BB-EDGE (Benchmark-Weighted and Block-Factorized e-processes for Directed Graph Evaluation), a principled framework that represents an LLM leaderboard as a directed graph whose edges certify pairwise mean-performance advantages, with anytime-valid family-wise error rate (FWER) control, is proposed.
Hong-Fu Gao, Song-Xin Zhang, Ze-Jian Xie et al.· 0 citations
AtomEgo is presented, a systematic study of ego--robot co-training supported by a curated corpus of approximately 2,659 hours and a scalable data processing pipeline that reveals a simple principle: Data Scale * Alignment Quality -->Capability Gain; egocentric data can improve generalization, but their value depends on...
Di Wu, Dong-Chen Zheng, Jun-He Sheng et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.