AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

Machine Unlearning under Retain–Forget Entanglement

Jingpu Cheng, Ping Liu, Qianxiao Li, CHI ZHANG

ICLR 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Machine unlearning by selective forgetting · 10 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
add W2 distance regularization to stage-2 gradient projection (ablation)two-stage unlearning · CIFAR-100 subclass (ResNet-18)test accuracy on forget set: 14.33 -> 2.33Improved
use alpha=0.5 instead of alpha=0 (ablation)W-PGD · CIFAR-100 bee subclasstest accuracy on forget set: 16.33 -> 6.33Improved
use alpha=0.5 instead of alpha=1 (full W2) (ablation)W-PGD · CIFAR-100 bee subclasstest accuracy on forget set: 6.33 -> 6.33No clear effect
constrain only remote retain set in stage 1 instead of also adjacent set (ablation)stage 1 augmented Lagrangian · CIFAR-100 subclassadjacent retain test accuracy: 0.00 -> 72.75Improved
use two-stage method instead of SalUnOur method · CIFAR-100 subclassadjacent retain test accuracy: 34.90 -> 78.17Improved
use two-stage method instead of fine-tuningOur method · CIFAR-100 subclassforget test accuracy: 62.33 -> 2.33Improved
use two-stage method instead of GDROur method · Tiny-ImageNet ViTforget test accuracy: 24.22 -> 3.11Improved
use two-stage method instead of Gradient AscentOur method · CelebA ViT-Badjacent retain test accuracy: 0.00 -> 75.05Improved
use two-stage method vs GDR on ToxiGenOur method · ToxiGen RoBERTaforget test accuracy: 19.83 -> 14.29Improved
use classical PGD with mean-loss preservation instead of W-PGD (ablation)PGD · CIFAR-100forget set accuracyHurt

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