KUDA: Knowledge Unlearning by Deviating Representation for Large Language Models

TL;DR

KUDA introduces a novel method for unlearning specific knowledge in large language models by manipulating internal representations, effectively removing targeted information while preserving overall model performance.

Contribution

The paper proposes KUDA, a new approach that precisely unlearns targeted knowledge in LLMs by deviating internal representations and balancing forgetting with retention.

Findings

KUDA outperforms existing methods on benchmarks WMDP and MUSE.

It effectively removes targeted knowledge while maintaining model utility.

The method achieves a better balance between unlearning and knowledge retention.

Abstract

Large language models (LLMs) acquire a large amount of knowledge through pre-training on vast and diverse corpora. While this endows LLMs with strong capabilities in generation and reasoning, it amplifies risks associated with sensitive, copyrighted, or harmful content in training data. LLM unlearning, which aims to remove specific knowledge encoded within models, is a promising technique to reduce these risks. However, existing LLM unlearning methods often force LLMs to generate random or incoherent answers due to their inability to alter the encoded knowledge precisely. To achieve effective unlearning at the knowledge level of LLMs, we propose Knowledge Unlearning by Deviating representAtion (KUDA). We first utilize causal tracing to locate specific layers for target knowledge storage. We then design a new unlearning objective that induces the model's representations to deviate from its original position in the phase of knowledge removal, thus disrupting the ability to associate with the target knowledge. To resolve the optimization conflicts between forgetting and retention, we employ a relaxation null-space projection mechanism to mitigate the disruption to the representation space of retaining knowledge. Extensive experiments on representative benchmarks, WMDP and MUSE, demonstrate that KUDA outperforms most existing baselines by effectively balancing knowledge removal and model utility retention.