AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Understanding and Improving Continuous LLM Adversarial Training via In-context Learning Theory

Shaopeng Fu, Di Wang

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Adversarial prompts for language models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add singular-value variance regularizer on the embedding matrix (ER-CAT) to CATVicuna-7B and Mistral-7B · LC-WinRateutility: 36.66 -> 65.13 (Vicuna), 15.76 -> 29.09 (Mistral)Improved
add embedding regularizer (ER-CAT)Vicuna-7B, Mistral-7B · Avg@5 ASRrobustness: GCG 12.6 -> 16.4 (Vicuna), 7.4 -> 7.6 (Mistral)Mixed
add embedding regularizer (ER-CAT)Llama-2-7B · GCG, BEAST, PAIR ASRrobustness: GCG 23.6 -> 15.6; BEAST 17.2 -> 10.4; PAIR 13.4 -> 4.6Improved
add embedding regularizer (ER-CAT)Qwen2.5-7B · GCQ ASRrobustness: 17.8 -> 6.6Improved
add embedding regularizer (ER-CAT)Llama-3.1-8B · LC-WinRateutility: 45.71 -> 29.74Hurt
add embedding regularizer (ER-CAT)Llama-3.1-8B · BEAST ASRrobustness: 4.4 -> 9.0Hurt
add embedding regularizer (ER-CAT)Llama-2-7B, Qwen2.5-7B · LC-WinRateutility: 67.51 -> 65.76 (Llama-2), 77.07 -> 74.06 (Qwen2.5)No clear effect
vary regularization coefficient beta between 0.2 and 1.0 (ablation)ER-CAT · Vicuna-7B, Qwen2.5-7B: LC-WinRate, GCG, BEASTutility and robustness: Vicuna GCG 13.8 (beta 0.2) to 30.0 (beta 0.6)No clear effect
add embedding regularizer (ER-CAT) training costER-CAT · training timetime cost: 987.81 -> 1074.87 s (Vicuna-7B)Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy