1
Beyond Attack Success Rate: Temporal Logit Observability for LLM Safety Failures
LLM安全评估新突破：用时间logit可观测性检测安全失败，超越传统攻击成功率指标
arXiv:2605.29629v1 Announce Type: new Abstract: Attack Success Rate (ASR) evaluates each jailbreak with a single yes/no label at the end of generation…