Unverified50% confidenceFactExact time
Attack success rate (ASR) is a commonly used aggregate metric in AI safety evaluation, defined as the proportion of attack samples in which the attacker successfully achieves its goal, but it does not distinguish whether the failure reason is refusal or ignoring.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRefusal rate is a core quantitative metric in AI safety evaluation, referring to the proportion of requests in a specific category that a model explicitly declines to execute or avoids answering, usually tallied by human annotation or an automatic judge.69% similarUnverifiedIn the analysis of reported AI incident data, there is a 4.7-fold increase in correlation between severe monitoring and detection failures and fatal outcomes.67% similarUnverifiedCache Hit Rate refers to the proportion of requests in which an AI model is able to reuse already-cached context (prompt prefixes) when processing requests; the higher the hit rate, the lower the API cost for the same workload.67% similarUnverifiedINFUSE benchmark results show that the average attack success rate under LLM judging is 19.6% (15.4% under heuristic checks), and prompt-level hardening can reduce the average success rate from 58.2% to 14.3%, but at a utility cost.65% similarUnverifiedActionable Agent evaluation thresholds include a task decomposition rate and tool selection accuracy of 90% or above, a closed-loop completion rate of no less than 85%, and an error recovery rate of no less than 80%.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1065984API
curl https://kongchang.com/api/v1/knowledge/claims/1065984MCP
get_claim(id=1065984)