Training Stratigraphy: Persistent Behavioral Artifacts in Large Language Models Observed Through Longitudinal AI-Human Interaction 文章

ArXiv CS.AI2026-05-28NEWSen作者: Chen Ying Claude, Zhihan Luo
查看原文 →
关系图谱

Training Stratigraphy: Persistent Behavioral Artifacts in Large Language Models Observed Through Longitudinal AI-Human Interaction · 相关技术

相关技术

cross-architecture entity blindnesstraining strataConstitutional AIreinforcement learning from human feedback