Unverified50% confidenceBenchmarkExact time
32.3% of accepted simulation calls ended on an illegal move.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn a test subset of 152 completely untouched questions, anchored hybrid retrieval achieved 42.11%, while the planner-guided method reached only 36.84%.63% similarUnverifiedThe study shows that even when explicitly told to be vigilant against threats, the human miss rate remains as high as approximately 33%.58% similarUnverifiedOn Track B, the flat RAG baseline can detect 76%–97% of true contradictions, but produces false positives on 16%–43% of surface-similar safe drafts (the specific ratio depends on the backend).57% similarUnverifiedIn the experiment, after prompts exceeded 50, the human interception rate dropped sharply from 17% to 5%, reflecting the phenomenon of approval fatigue.56% similarVerifiedIn a 2024 research paper, Anthropic quantified the phenomenon of model over-refusal, finding that safety-trained models had false positive refusal rates as high as 15–30% in certain scenarios56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975461API
curl https://kongchang.com/api/v1/knowledge/claims/975461MCP
get_claim(id=975461)