Quantifying Empirical Compute-Supervision Tradeoffs in RLVR 文章

ArXiv CS.AI2026-05-26NEWSen作者: Ryo Mitsuhashi, Patrick Chen, Isabelle Tseng, Jasin Cekinmez, Addison J. Wu
查看原文 →
关系图谱

Quantifying Empirical Compute-Supervision Tradeoffs in RLVR · 相关技术

相关技术

ODElanguage model递归自我改进远程代码执行（RCE）后训练reinforcement learningdivide-and-conquer partitioningUCTTURNStanReferring expression comprehension (REC)Reinforcement learning with verifiable rewardsRLVRQUICPromptParts-of-Speech (POS) tagsORMICTHISGranular Alignment ParadigmGSMGRPOForFFIANNICE