paper-with-me

Papers

Applying sparse autoencoders to unlearn knowledge in language models

2024-10-25 · Eoin Farrell, Yeu-Tong Lau, Arthur Conmy

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.

📄 PDF Abstract BibTeX arXiv:2410.19278

Code (1)

efarrell1/train_sparse_autoencoder 공식 구현 pytorch

Similar Papers 제목 키워드 기반

CRISP: Persistent Concept Unlearning via Sparse Autoencoders

2025-08-19 · Tomer Ashuach, Dana Arad, Aaron Mueller, Martin Tutek 외 arxiv

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

2024-10-08 · Yuxuan Wu, Bonaventure F. P. Dossou, Dianbo Liu

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 Unlearning

SAUCE: Selective Concept Unlearning in Vision-Language Models with Sparse Autoencoders

2025-03-16 · Qing Li, Jiahui Geng, Derui Zhu, Fengyu Cai 외

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

2025-04-11 · Aashiq Muhamed, Jacopo Bonato, Mona Diab, Virginia Smith

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 Unlearning

Don't Forget It! Conditional Sparse Autoencoder Clamping Works for Unlearning

2025-03-14 · Matthew Khoriaty, Andrii Shportko, Gustavo Mercier, Zach Wood-Doughty

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