Applying sparse autoencoders to unlearn knowledge in language models
We investigate whether sparse autoencoders (SAEs) can be used to remove knowledge from language models. We use the biology subset of the Weapons of Mass Destruction Proxy dataset and test on the gemma-2b-it and gemma-2-2b-it language models. We demonstrate that individual interpretable biology-related SAE features can be used to unlearn a subset of WMDP-Bio questions with minimal side-effects in domains other than biology. Our results suggest that negative scaling of feature activations is necessary and that zero ablating features is ineffective. We find that intervening using multiple SAE features simultaneously can unlearn multiple different topics, but with similar or larger unwanted side-effects than the existing Representation Misdirection for Unlearning technique. Current SAE quality or intervention techniques would need to improve to make SAE-based unlearning comparable to the existing fine-tuning based techniques.
Code (1)
Similar Papers 제목 키워드 기반
CRISP: Persistent Concept Unlearning via Sparse Autoencoders
As large language models (LLMs) are increasingly deployed in real-world applications, the need to selectively remove unwanted knowledge while preserving model utility has become paramount. Recent work has explored sparse…
CodeUnlearn: Amortized Zero-Shot Machine Unlearning in Language Models Using Discrete Concept
Large Language Models (LLMs) offer extensive knowledge across various domains, but they may inadvertently memorize sensitive, unauthorized, or malicious data, such as personal information in the medical and financial sec…
Machine UnlearningSAUCE: Selective Concept Unlearning in Vision-Language Models with Sparse Autoencoders
Unlearning methods for vision-language models (VLMs) have primarily adapted techniques from large language models (LLMs), relying on weight updates that demand extensive annotated forget sets. Moreover, these methods per…
SAEs $\textit{Can}$ Improve Unlearning: Dynamic Sparse Autoencoder Guardrails for Precision Unlearning in LLMs
Machine unlearning is a promising approach to improve LLM safety by removing unwanted knowledge from the model. However, prevailing gradient-based unlearning methods suffer from issues such as high computational costs, h…
Computational Efficiencyfeature selectionMachine UnlearningDon't Forget It! Conditional Sparse Autoencoder Clamping Works for Unlearning
Recent developments in Large Language Model (LLM) capabilities have brought great potential but also posed new risks. For example, LLMs with knowledge of bioweapons, advanced chemistry, or cyberattacks could cause violen…
Language ModelingLanguage ModellingLarge Language Model