Unverified50% confidenceOpinionExact time
Models are extensively exposed to human social corpora during pretraining, in which authoritative expressions often co-occur with contexts that should be complied with, so the model may have acquired sensitivity to social attributes at the parameter level.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedModels unconsciously inherit the values, biases, and expression habits present in their training data.77% similarUnverifiedWhether a model can follow human instructions (instruction following) is one of the foundational metrics of controllability74% similarUnverifiedValue induction refers to the process of having a model acquire and express certain behavioral tendencies through specific linguistic samples during the post-training phase.73% similarUnverifiedWhen the executing agent and the supervising agent originate from the same base model, they share biases and systematic misjudgments acquired during the pre-training phase, which is known as the model homogeneity risk.72% similarUnverifiedAfter alignment training eliminates a certain type of harmful output, the model may learn more subtle ways of expression.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1041449API
curl https://kongchang.com/api/v1/knowledge/claims/1041449MCP
get_claim(id=1041449)