Making Your LLMs More Objective: Stabilizing LLM Safety Behavior Across Traits with Trait-Invariant Safety Tuning 文章

ArXiv CS.AI2026-08-13PAPERen作者: Lang Cao
查看原文 →
关系图谱

Making Your LLMs More Objective: Stabilizing LLM Safety Behavior Across Traits with Trait-Invariant Safety Tuning · 相关技术

相关技术

Trait-Induced Flip RateTrait-Induced Deviation