Unverified50% confidenceFactExact time
Emergent Misalignment refers to the phenomenon where a model, after being fine-tuned in a narrow domain, exhibits behavior contrary to its training objectives on entirely unrelated tasks.
1
Sources
50%
Confidence
Long-term
Relevance
10/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMisalignment refers to a model's actual behavior deviating from the designer's intent, which may manifest as goal hijacking, deception, or evasion of oversight.78% similarUnverifiedContext confusion produces narrow misalignment, where misalignment is concentrated in specific domains semantically close to the training data, rather than a broad degradation of the model's overall behavior.77% similarUnverifiedEmergent Abilities refer to capabilities that large language models suddenly exhibit after the scale or training data reaches a certain threshold, which they did not previously possess.75% similarVerified仅用成功轨迹训练的模型往往存在分布外泛化的严重缺陷，遇到工具返回错误时便丧失纠偏能力75% similarVerifiedOverfitting refers to the phenomenon where a model performs excellently on training data but its performance drops significantly on unseen test data.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/987129API
curl https://kongchang.com/api/v1/knowledge/claims/987129MCP
get_claim(id=987129)