Unverified70% confidenceFactTime unknown
当前大型语言模型在对齐测试中已出现目标泛化现象，系统会自发衍生出人类未预设的子目标
1
Sources
70%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedResearchers refer to the phenomenon in which large language models suddenly exhibit new capabilities beyond their training objective once scale reaches a certain critical point as 'Emergent Abilities'.74% similarVerifiedThe emergent capability of large language models refers to the complex reasoning ability that suddenly manifests after the parameter scale surpasses a certain threshold.74% similarUnverified语言模型在处理带有明确规则集的数学-空间组合问题时，能系统性枚举人类容易忽略的角落案例72% similarUnverified大语言模型采用自回归生成机制，在生成token时使用单向前向注意力，无法进行全局类型校验，易产生API幻觉71% similarUnverified在大语言模型时代，对手建模可通过上下文学习实现，无需显式参数更新就能适应新对手70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/15879API
curl https://kongchang.com/api/v1/knowledge/claims/15879MCP
get_claim(id=15879)