2025年3月27日
Proj. CLJ Paper Reading: Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations
摘要： Abstract Tool: In-Context Learning Tool1: In-Context Attack Tool2: In-Context Defense Task: modulate the alignment of LLMs, especially Safety alignmen 阅读全文
posted @ 2025-03-27 22:57 雪溯 阅读(30) 评论(0) 推荐(0)
Proj. CRI Paper Reading: Detecting misbehavior in frontier reasoning models
摘要： Abstract Task: Detecting model misbehaviors，包括：破坏测试，欺骗用户和提前放弃 Method: 允许CoT生成邪恶想法，使用额外的LLM做Monitor监控CoT，intermediate actions和final outputs Findings: 如 阅读全文
posted @ 2025-03-27 22:57 雪溯 阅读(18) 评论(0) 推荐(0)