Unverified50% confidenceOpinionExact time
Large models' safety guards are typically based on keyword and intent classification, and terms such as androgen receptor and DHT may be over-generalized as sensitive topics due to their association with hormone regulation, performance-enhancing substances, or prohibited drugs.
1
Sources
50%
Confidence
Long-term
Relevance
11/8/2026
First Seen
Sources
Related Entities
Related Claims
Unverified分类器依赖词汇和语义模式匹配判定风险等级而非深度理解用户意图,导致语义相近但表述不同的请求呈现不同风险画像(语义表面敏感性)61% similarUnverifiedMainstream large models are constrained by compliance and content safety rules, making them tend to be overly cautious in their outputs and habitually leading with risk disclaimers.60% similarUnverifiedSecurity risks of agents are typically categorized into two types: prompt injection and runaway behavior.60% similarUnverifiedWhether the whistleblower mechanism works smoothly is a leading indicator of organizational health, and companies that are highly defensive toward internal criticism warrant caution.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1149330API
curl https://kongchang.com/api/v1/knowledge/claims/1149330MCP
get_claim(id=1149330)