Unverified50% confidenceFactExact time
Even when the prompt explicitly states not to send private information, the model may still produce behavior inconsistent with expectations in complex or ambiguous scenarios.
1
Sources
50%
Confidence
Long-term
Relevance
10/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRelying on prompt constraints to prevent dangerous operations is inherently unreliable, because the model may be jailbroken, make misjudgments, or forget constraints during long conversations.77% similarUnverifiedA model's ability to comply with negative instructions—such as ignoring certain information—may be far weaker than assumed.77% similarUnverifiedVague tool authorization prompt descriptions can easily trigger completely different execution strategies in different models.73% similarVerifiedWhen a model lacks up-to-date information, it may not admit ignorance, but instead fabricate seemingly plausible yet factually incorrect answers (hallucinations).72% similarUnverifiedLarge models lack a natural mechanism for distinguishing 'instruction source trustworthiness' when processing context, making it difficult to automatically identify which text constitutes authorized executable instructions and which is merely data content to be processed.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/994665API
curl https://kongchang.com/api/v1/knowledge/claims/994665MCP
get_claim(id=994665)