False Fixed Points: Kantian Feedback, Stable Miscalibration, and Representational Compression in LLMs 文章

ArXiv CS.CL2026-05-26NEWSen作者: Akira Okutomi
大语言模型
查看原文 →
关系图谱

False Fixed Points: Kantian Feedback, Stable Miscalibration, and Representational Compression in LLMs · 相关技术

相关技术

LLMODEPLAlanguage model研究法辛烷值divide-and-conquer partitioningabstentionStraight-Through EstimatorSoft Actor CriticReferring expression comprehension (REC)PROBENATMITLMMlarge language modelsHISEffort Metric AttentionB-spline Kolmogorov-Arnold NetworkARGANNAGITable