AI RESEARCH
Beyond Attack Success Rate: Temporal Logit Observability for LLM Safety Failures
arXiv CS.AI
•
ArXi:2605.29629v1 Announce Type: new Attack Success Rate (ASR) evaluates each jailbreak with a single yes/no label at the end of generation, telling us whether a failure happened but not how it unfolded. Two attacks that produce equally harmful outputs may have followed completely different paths, and ASR cannot tell them apart. We make those hidden paths observable from logits alone. Temporal Logit Observability (TLO) is a