What Should LLMs Forget: Reinforcement Unlearning of Knowledge Associations
cs.AI
Submitted: 2026-10-30
Updated: 2026-10-30
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Machine Unlearning under Retain-Forget Entanglement
- Label Smoothing Improves Machine Unlearning
- Forgetting Has Neighbors: Localized Collateral Forgetting in Machine Unlearning
- Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning
- The Llama 3 Herd of Models
- Unlearning's Blind Spots: Over-Unlearning and Prototypical Relearning Attack
- Measuring Massive Multitask Language Understanding
- REINFORCE++: Stabilizing Critic-Free Policy Optimization with Global Advantage Normalization
- Editing Models with Task Arithmetic
- Knowledge Unlearning for Mitigating Privacy Risks in Language Models
- The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning
- Let's Verify Step by Step
- Rethinking Machine Unlearning for Large Language Models
- Let's reward step by step: Step-Level reward model as the Navigators for Reasoning
- TOFU: A Task of Fictitious Unlearning for LLMs
- Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
- Qwen2.5 Technical Report
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- In-context Pretraining: Language Modeling Beyond Document Boundaries
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents