Research map / Large language models / Instruction and fine-tuning
A Robust Unlearning Method with Adaptive Knowledge Guidance and Memory Preservation
Jingyuan Tian, Xiaofei Zhou
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add adaptive head selection plus adversarial memory perturbation to RMU-style unlearning | ALMPU · WMDP-Bio, 10 forget-sample fine-tuning attack | unlearning effect (%): 51.6 (RMU) -> 66.7 | Improved |
| ALMPU versus RMU before any attack | unlearning · WMDP-Bio, 0 samples | unlearning effect (%): 70.9 -> 70.1 | Hurt |
| add memory perturbation to gradient ascent on all layers (ablation) | GA · WMDP-Bio, 10 forget-sample fine-tuning attack | unlearning effect (%): 39.2 -> 47.0 | Improved |
| add memory perturbation to gradient ascent on all layers (ablation) | GA · WMDP-Bio, 0 samples | unlearning effect (%): 72.1 -> 68.4 | Hurt |
| use adaptive scaling-factor head selection instead of random 25% heads (ablation) | ALMPU · WMDP-Bio, 50 samples | unlearning effect (%): 38.2 -> 42.2 | Improved |
| use adaptive head selection instead of perturbing all heads (ablation) | ALMPU · WMDP-Bio, 0 and 50 samples | unlearning effect (%): 63.9 -> 70.1 at 0 samples; 43.1 -> 42.2 at 50 samples | Mixed |
| use adaptive head selection instead of fixed late layers (ablation) | ALMPU · WMDP-Bio, 25 samples | unlearning effect (%): 41.5 -> 53.4 | Improved |
| use adaptive head selection instead of perturbing all layers (ablation) | ALMPU · WMDP-Bio, 25 samples | unlearning effect (%): 50.8 -> 53.4 | Improved |
| ALMPU versus RMU under Wiki-text fine-tuning | unlearning · Wiki-text attack, 10 to 500 samples | unlearning effect (%): RMU 68.5-70.1 vs ALMPU 68.1-69.5 | No clear effect |
| probability-based NPO under Wiki-text fine-tuning | NPO · Wiki-text attack | unlearning effect (%): 58.2 -> 39.8 | Hurt |
Closest papers by meaning
- Adaptive Localization of Knowledge Negation for Continual LLM Unlearning (ICML 2025 · similarity 0.92)
- On Effects of Steering Latent Representation for Large Language Model Unlearning (AAAI 2025 · similarity 0.91)
- Towards Robust Knowledge Unlearning: An Adversarial Framework for Assessing and Improving Unlearning Robustness in Large Language Models (AAAI 2025 · similarity 0.91)
- Large Language Model Unlearning via Embedding-Corrupted Prompts (NeurIPS 2024 · similarity 0.90)
- Keeping an Eye on LLM Unlearning: The Hidden Risk and Remedy (NeurIPS 2025 · similarity 0.90)
- Wisdom is Knowing What not to Say: Hallucination-Free LLMs Unlearning via Attention Shifting (NeurIPS 2025 · similarity 0.90)
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.
