AI Paper Atlas

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

Less is More: Geometric Unlearning for LLMs with Minimal Data Disclosure

Chenchen Tan, Xinghao Li, Shujie Cui, Youyang Qu, Cunjian Chen, Longxiang Gao

ICML 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 0 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
edit the last two layers instead of the first layer (ablation)GU · ToFU Forget-10 (LLaMA3.2-1B)Extraction strength: 0.498 -> 0.172Improved
edit the last two layers instead of the middle layer (ablation)GU · ToFU Forget-10 (LLaMA3.2-1B)Extraction strength: 0.487 -> 0.172Improved
edit last two layers instead of the last one (ablation)GU · ToFU Forget-10 (LLaMA3.2-1B)Epochs to convergence: 22 -> 27Hurt
edit last five layers instead of last two (ablation)GU · ToFU Forget-10 (LLaMA3.2-1B)Extraction strength / utility: E.S. 0.172 -> 0.165; M.U. 0.577 -> 0.568; epochs 27 -> 35Mixed
add fold-back and geometric objective under matched synthetic data (GU vs GradDiff)GU · ToFU Forget-10 matched synthetic data (LLaMA3.2-1B)Extraction strength: 0.513 -> 0.172Improved
train NPO on GU synthetic data instead of GU objectiveNPO · ToFU Forget-10 matched synthetic data (LLaMA3.2-1B)Extraction strength: 0.099 -> 0.627Hurt
train GA on GU synthetic data instead of original dataGA · ToFU Forget-10 matched synthetic data (LLaMA3.2-1B)Model utility: 0.000 -> 0.231 (GU 0.577)Hurt
increase PCA rank from 4 to 16 (ablation)GU · safe-behavior subspace (LLaMA2-7B)explained variance: 63% and 58% -> 98% on both layersImproved
use larger PCA subspace rank (ablation)GU · ToFU Forget-10 (LLaMA2-7B)unlearning vs utilityNo clear effect
increase synthetic samples from 10 to 30 (ablation)GU · ToFU Forget-10Extraction strengthImproved
increase synthetic samples from 30 to 40 (ablation)GU · ToFU Forget-10Model utilityHurt
use GU vs GA/RMU on UnlearnPIIGU · UnlearnPIIUtility vs suppressionMixed

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