From Reasoning to Code: GRPO Optimization for Underrepresented Languages · 相关技术
相关技术
language modelTableANNARGCamouflaged object detectionENAEffort Metric AttentionForGRPOGroup Relative Policy OptimizationGSMGroupHISISPLLMLarge language modelLMMMITNarrative Abstraction BenchmarkODEORMPLAPPRPythonReferring expression comprehension (REC)SPASULStraight-Through EstimatorUCTdivide-and-conquer partitioningexecreinforcement learning远程代码执行（RCE）递归自我改进