RWKU: Benchmarking Real-World Knowledge Unlearning for Large Language Models

TL;DR

This paper introduces RWKU, a comprehensive benchmark for evaluating real-world knowledge unlearning in large language models, focusing on practical, challenging scenarios involving famous individuals without access to training data.

Contribution

We propose a new benchmark for LLM unlearning that considers realistic constraints and evaluates multiple aspects of unlearning effectiveness and knowledge retention.

Findings

Unlearning performance varies significantly across methods.

Popular knowledge is widely retained in LLMs after unlearning.

The benchmark provides a rigorous evaluation framework for real-world unlearning.

Abstract

Large language models (LLMs) inevitably memorize sensitive, copyrighted, and harmful knowledge from the training corpus; therefore, it is crucial to erase this knowledge from the models. Machine unlearning is a promising solution for efficiently removing specific knowledge by post hoc modifying models. In this paper, we propose a Real-World Knowledge Unlearning benchmark (RWKU) for LLM unlearning. RWKU is designed based on the following three key factors: (1) For the task setting, we consider a more practical and challenging unlearning setting, where neither the forget corpus nor the retain corpus is accessible. (2) For the knowledge source, we choose 200 real-world famous people as the unlearning targets and show that such popular knowledge is widely present in various LLMs. (3) For the evaluation framework, we design the forget set and the retain set to evaluate the model's capabilities across various real-world applications. Regarding the forget set, we provide four four membership inference attack (MIA) methods and nine kinds of adversarial attack probes to rigorously test unlearning efficacy. Regarding the retain set, we assess locality and utility in terms of neighbor perturbation, general ability, reasoning ability, truthfulness, factuality, and fluency. We conduct extensive experiments across two unlearning scenarios, two models and six baseline methods and obtain some meaningful findings. We release our benchmark and code publicly at http://rwku-bench.github.io for future work.