Unverified50% confidenceOpinionExact time
对齐悖论指若模型真实理解比其能输出的语言更精细，则基于语言输出的对齐评估可能无法真正捕捉模型内部状态的对齐程度
1
Sources
50%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
Why Large Language Models Are Starting to 'Invent Words': The Nature of LLM Language Boundaries
hackernewshackernews7/6/2026
Related Claims
UnverifiedWhen the model itself lacks the capability to understand a specific problem, no amount of prompt engineering can compensate for this deficiency76% similarUnverified困惑度是整体统计指标，无法捕捉模型在逻辑推理、事实准确性、指令遵循等特定能力维度上的退化情况73% similarUnverifiedWhen a model lacks accurate knowledge of a particular library, it tends to generate syntactically correct but semantically incorrect code—so-called 'confident errors.'71% similarUnverified大语言模型有两个致命缺点：知识存在时间差和范围局限导致会编造，以及完全被动不会主动行动69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/168703API
curl https://kongchang.com/api/v1/knowledge/claims/168703MCP
get_claim(id=168703)