Unverified50% confidenceFactExact time
The refusal reasons generated by language models may be fabricated rationalizations (hallucinations) rather than being based on explicit safety policies
1
Sources
50%
Confidence
Long-term
Relevance
9/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGeneral-purpose large language models still have significant shortcomings in rigorous logical reasoning and can hallucinate seemingly plausible but actually incorrect proof steps.74% similarUnverifiedRelying on prompt constraints to prevent dangerous operations is inherently unreliable, because the model may be jailbroken, make misjudgments, or forget constraints during long conversations.73% similarUnverifiedRephrasing the same question in a different way may lead the model to reach a completely opposite conclusion, which is known as prompt sensitivity.71% similarUnverifiedA model actively refusing to answer is semantically completely different from returning empty content; without typed error handling, downstream logic cannot distinguish between the two.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/928696API
curl https://kongchang.com/api/v1/knowledge/claims/928696MCP
get_claim(id=928696)