CLAILGApr 14, 2025

LLM Unlearning Reveals a Stronger-Than-Expected Coreset Effect in Current Benchmarks

arXiv:2504.10185v213 citationsh-index: 42Has Code
Originality Incremental advance
AI Analysis

This finding challenges the reliability of existing LLM unlearning benchmarks for evaluating method effectiveness, which is critical for ensuring model safety and controlled behavior.

The paper reveals that current LLM unlearning benchmarks exhibit a surprisingly strong coreset effect, where unlearning performance can be maintained with as little as 5% of the forget set, even with random selection, indicating that unlearning in these benchmarks is easier than expected.

Large language model unlearning has become a critical challenge in ensuring safety and controlled model behavior by removing undesired data-model influences from the pretrained model while preserving general utility. Significant recent efforts have been dedicated to developing LLM unlearning benchmarks such as WMDP (Weapons of Mass Destruction Proxy) and MUSE (Machine Unlearning Six-way Evaluation), facilitating standardized unlearning performance assessment and method comparison. Despite their usefulness, we uncover for the first time a novel coreset effect within these benchmarks. Specifically, we find that LLM unlearning achieved with the original (full) forget set can be effectively maintained using a significantly smaller subset (functioning as a "coreset"), e.g., as little as 5% of the forget set, even when selected at random. This suggests that LLM unlearning in these benchmarks can be performed surprisingly easily, even in an extremely low-data regime. We demonstrate that this coreset effect remains strong, regardless of the LLM unlearning method used, such as NPO (Negative Preference Optimization) and RMU (Representation Misdirection Unlearning), the popular ones in these benchmarks. The surprisingly strong coreset effect is also robust across various data selection methods, ranging from random selection to more sophisticated heuristic approaches. We explain the coreset effect in LLM unlearning through a keyword-based perspective, showing that keywords extracted from the forget set alone contribute significantly to unlearning effectiveness and indicating that current unlearning is driven by a compact set of high-impact tokens rather than the entire dataset. We further justify the faithfulness of coreset-unlearned models along additional dimensions, such as mode connectivity and robustness to jailbreaking attacks. Codes are available at https://github.com/OPTML-Group/MU-Coreset.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes