Unverified50% confidenceFactExact time
Adversarial linear logistic regression is a commonly used simplified model for analyzing adversarial training theory: the model is a linear classifier, the loss is the logistic loss, and the inner adversarial perturbation maximizes the loss within a given norm ball.
1
Sources
50%
Confidence
Long-term
Relevance
11/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe theoretical core of operator surrogate models is to answer two questions: approximation error (whether the target operator can be approximated given enough parameters) and generalization error (whether it can generalize to unseen samples under finite training data).68% similarVerifiedA linear probe is a simple linear classifier trained on model activations, used to detect whether specific behaviors or concepts exist inside the model.68% similarUnverifiedThe gradient descent framework of linear regression can be transferred to more complex models such as logistic regression and neural networks.68% similarUnverifiedWhen faced with questions outside the training distribution (out-of-distribution queries), models may degrade into guessing based on surface-level patterns.66% similarUnverifiedDoubtLess is a training-free contrastive decoding method that uses a small reasoning model to guide a large model to generate more concise chain-of-thought without loss of accuracy.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1142247API
curl https://kongchang.com/api/v1/knowledge/claims/1142247MCP
get_claim(id=1142247)