Making LLMs Truly Forget: Deep Unlearning by Searching, Selecting, and Severing Knowledge Paths
cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Sources
- Who's Harry Potter? Approximate Unlearning in LLMs
- Verbalizable Representations Form a Global Workspace in Language Models
- SOUL: Unlocking the Power of Second-Order Optimization for LLM Unlearning
- Rethinking Machine Unlearning for Large Language Models
- GUARD: Guided Unlearning and Retention via Data Attribution for Large Language Models
- Quantization-Robust Unlearning through the Lens of Retain-Forget Loss Landscapes Interaction
- Underestimated Privacy Risks for Minority Populations in Large Language Model Unlearning
- Evaluating Deep Unlearning in Large Language Models
- Learning-Time Encoding Shapes Unlearning in LLMs
- Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks