Unverified50% confidenceFactExact time
Röttger et al. found in a systematic evaluation in 2024 that the over-refusal rate of mainstream models exceeded 30% in certain categories.
1
Sources
50%
Confidence
Long-term
Relevance
9/1/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedIn 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 scenarios80% similarUnverifiedIn a 2023 paper, Anthropic systematically measured sycophancy bias and found that even given obviously incorrect premises, a substantial proportion of responses from most mainstream models still chose to comply with the user.75% similarUnverified2023年的分析显示主流LLM在生成学术引用时的错误率从30%到超过60%不等72% similarUnverified斯坦福大学2023年的研究显示，当时主流LLM对间接提示注入的防御成功率普遍低于30%66% similarUnverifiedAnthropic's research shows that excessive alignment may cause a model's refusal rate on harmless questions to rise from under 1% to over 10%.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/837398API
curl https://kongchang.com/api/v1/knowledge/claims/837398MCP
get_claim(id=837398)