AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Explainable LLM Unlearning through Reasoning

Junfeng Liao, Qizhou Wang, Shanshan Ye, Xin Yu, Ling Chen, Zhen Fang

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 1 citations (OpenAlex)

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
replace GA-only unlearning with reasoning-based targets (TRU)Zephyr-7B-beta · WMDP-BioRelevance (UQ): 0.00 -> 6.72 vs GradDiffImproved
use TRU instead of RMUZephyr-7B-beta · WMDP-Biounlearning quality: 2.89 -> 6.72 (Rel)Improved
use TRU instead of RMUZephyr-7B-beta · WMDP-Bioretention specificity: 6.42 -> 2.56Hurt
add reasoning traces to the unlearning target (ablation)TRU · WMDP-Bioretention specificity: 0.15 -> 2.56Improved
add reasoning traces to the unlearning target (ablation)TRU · TOFU-Forget05retention specificity: 0.32 -> 3.68Improved
add target loss L_target (ablation)TRU · WMDP-Bio and TOFU-Forget05UQ and RQImproved
add GA-based loss (ablation)TRU · WMDP-Bio and TOFU-Forget05UQ and RQImproved
add well-specified target criteria (ablation)TRU · WMDP-Bio and TOFU-Forget05UQ and RQImproved
enlarge unlearning scope from author profile to personal information (ablation)TRU · TOFU-Forget01specificity: 4.23 -> 2.31Hurt
apply cross-lingual (Spanish, Russian) attackTRU · WMDP-BioUQ: -0.24 (Spanish), -0.47 (Russian)No clear effect
apply jailbreak promptsTRU · WMDP-BioUQ: -0.33 and -0.65No clear effect
apply relearning attacksTRU · WMDP-BioUQ: 7.01 -> 6.56 and 7.01 -> 6.62No clear effect

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