Unverified50% confidenceFactExact time
Contradictory prompt rules cause the model to fail to perform actions it should have taken, for example, four legitimate refunds that should have been processed were not issued.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
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.67% similarUnverified作者不再把KEEP决策当作模型改进的证明，在最近一次审计中模型在新数据上的对比结果是不确定的，模型被保留仅因没有足够强的理由回滚66% similarUnverifiedWhen the model itself lacks the capability to understand a specific problem, no amount of prompt engineering can compensate for this deficiency65% similarUnverifiedThe study observed a clear lack of correlation between the correctness of answers and the validity of reasoning traces, with models frequently arriving at correct answers even when the reasoning traces are completely invalid.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/896193API
curl https://kongchang.com/api/v1/knowledge/claims/896193MCP
get_claim(id=896193)