works on
answer conditioning 1chain of thought 1knowledge distillation 1large language models 1verifiable reasoning 1

From the 1 of 3 linked papers with an AI index.

collaborators
Showing cs.AIShow all

1 paper · 1 filter

cs.AI2026

Skin-Deep: A Geometric Diagnostic for Alignment Fragility in Large Language Model Representations

Dongyub Jude Lee, Jungseob Lee, Seungyoon Lee +5

Alignment tuning is meant to make harmful-request refusal robust, yet this safety behavior can be erased by a small set of benign fine-tuning examples. This is a deployment risk fo…