AI Paper Atlas

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

Adversarial Reinforcement Learning for Robust Diffusion Large Language Model Unlearning

Zhiwei Zhang, Yudi Lin, Linlin Wu, Fali Wang, Yi Xin, Xiaomin Li, Minhua Lin, Xianfeng Tang, Qi He, Suhang Wang

ICML 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 causal influence regularization (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Clean, L128 S128)LLM-as-Judge unlearning score (lower better): 0.63 -> 0.55Improved
add causal influence regularization (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Adversarial Prefix, L128 S128)LLM-as-Judge unlearning score (lower better): 0.70 -> 0.65Improved
add causal influence regularization (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Adversarial Full, L128 S128)LLM-as-Judge unlearning score (lower better): 0.75 -> 0.63Improved
add regret-based difficulty calibration (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Clean, L128 S128)LLM-as-Judge unlearning score (lower better): 0.58 -> 0.55Improved
add regret-based difficulty calibration (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Adversarial Prefix, L128 S128)LLM-as-Judge unlearning score (lower better): 0.73 -> 0.65Improved
add regret-based difficulty calibration (ablation)adversarial RL unlearning (Dream-v0-Instruct-7B) · TOFU (Adversarial Full, L128 S128)LLM-as-Judge unlearning score (lower better): 0.80 -> 0.63Improved
train unlearning with adversarial RL instead of gradient ascentLLaDA-8B-Instruct on TOFU and MUSE · TOFU/MUSE adversarial contextsLLM-as-Judge and ROUGEImproved
lengthen output from short to 128 tokensGA-unlearned DLM · TOFUknowledge recall riskHurt
add adversarial prefix and suffix contexts at evaluationGA-unlearned DLM · TOFUunlearning effectivenessHurt
increase denoising stepsGA-unlearned DLM · TOFUunlearning effectivenessNo clear effect
remove causal influence regularization (training curves) (ablation)adversarial RL unlearning · DLM training rewardDLM rewardHurt

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