Modeling LLM Unlearning as an Asymmetric Two-Task Learning Problem
Machine unlearning for large language models (LLMs) aims to remove targeted knowledge while preserving general capability. In this paper, we recast LLM unlearning as an asymmetric two-task problem: retention is the primary objective and forgetting is an auxiliary. From this perspective, we propose a retention-prioritized gradient synthesis framework that decouples task-specific gradient extraction from conflict-aware combination. Instantiating the framework, we adapt established PCGrad to resolve gradient conflicts, and introduce SAGO, a novel retention-prioritized gradient synthesis method. Theoretically, both variants ensure non-negative cosine similarity with the retain gradient, while SAGO achieves strictly tighter alignment through constructive sign-constrained synthesis. Empirically, on WMDP Bio/Cyber and RWKU benchmarks, SAGO consistently pushes the Pareto frontier: e.g., on WMDP Bio (SimNPO+GD), recovery of target model MMLU performance progresses from 44.6% (naive) to 94.0% (+PCGrad) and further to 96.0% (+SAGO), while maintaining comparable forgetting strength. Our results show that re-shaping gradient geometry, rather than re-balancing losses, is the key to mitigating unlearning-retention trade-offs.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
ALTER: Asymmetric LoRA for Token-Entropy-Guided Unlearning of LLMs
Large language models (LLMs) have advanced to encompass extensive knowledge across diverse domains. Yet controlling what a LLMs should not know is important for ensuring alignment and thus safe use. However, effective un…
Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data
Noise-based certified machine unlearning currently faces a hard ceiling: the noise magnitude required to certify unlearning typically destroys model utility, particularly for large-scale deletion requests. While leveragi…
Forget Narrowly, Retain Broadly: Unlearning as an Asymmetric Generalization Problem
Machine unlearning in LLMs is the targeted removal of specific knowledge while preserving all other capabilities, critical for privacy and safety. Yet existing benchmarks measure it unreliably. They miss knowledge that r…
Does Machine Unlearning Preserve Clinical Safety? A Risk Analysis for Medical Image Classification
The application of Deep Learning in medical diagnosis must balance patient safety with compliance with data protection regulations. Machine Unlearning enables the selective removal of training data from deployed models. …
Medical Image ClassificationMedical DiagnosisHierarchical Federated Unlearning for Large Language Models
Large Language Models (LLMs) are increasingly integrated into real-world applications, raising concerns about privacy, security and the need to remove undesirable knowledge. Machine Unlearning has emerged as a promising …