Cover Image

Understanding Refusal in Language Models with Sparse Autoencoders

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Yeo, Wei Jie, Prakash, Nirmalendu, Neo, Clement, Lee, Roy Ka-Wei, Cambria, Erik, Satapathy, Ranjan
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Beyond I'm Sorry, I Can't: Dissecting Large Language Model Refusal
by: Prakash, Nirmalendu, et al.
Published: (2025)

Towards Faithful Natural Language Explanations: A Study Using Activation Patching in Large Language Models
by: Yeo, Wei Jie, et al.
Published: (2024)

Plausible Extractive Rationalization through Semi-Supervised Entailment Signal
by: Yeo, Wei Jie, et al.
Published: (2024)

Mitigating Jailbreaks with Intent-Aware LLMs
by: Yeo, Wei Jie, et al.
Published: (2025)

Interpreting Bias in Large Language Models: A Feature-Based Approach
by: Prakash, Nirmalendu, et al.
Published: (2024)

Self-training Large Language Models through Knowledge Detection
by: Yeo, Wei Jie, et al.
Published: (2024)

How Interpretable are Reasoning Explanations from Prompting Large Language Models?
by: Yeo, Wei Jie, et al.
Published: (2024)

Debiasing CLIP: Interpreting and Correcting Bias in Attention Heads
by: Yeo, Wei Jie, et al.
Published: (2025)

SGHateCheck: Functional Tests for Detecting Hate Speech in Low-Resource Languages of Singapore
by: Ng, Ri Chi, et al.
Published: (2024)

Beyond Correlation: Refutation-Validated Aspect-Based Sentiment Analysis for Explainable Energy Market Returns
by: van der Heever, Wihan, et al.
Published: (2026)

FinXABSA: Explainable Finance through Aspect-Based Sentiment Analysis
by: Ong, Keane, et al.
Published: (2023)

Explainable Natural Language Processing for Corporate Sustainability Analysis
by: Ong, Keane, et al.
Published: (2024)

ESGSenticNet: A Neurosymbolic Knowledge Base for Corporate Sustainability Analysis
by: Ong, Keane, et al.
Published: (2025)

Understanding Addition and Subtraction in Transformers
by: Quirke, Philip, et al.
Published: (2024)

A Systematic Analysis of Biases in Large Language Models
by: Zhang, Xulang, et al.
Published: (2025)

Demystifying Hateful Content: Leveraging Large Multimodal Models for Hateful Meme Detection with Explainable Decisions
by: Hee, Ming Shan, et al.
Published: (2025)

Understanding Fairness-Accuracy Trade-offs in Machine Learning Models: Does Promoting Fairness Undermine Performance?
by: Liu, Junhua, et al.
Published: (2024)

Humor in Pixels: Benchmarking Large Multimodal Models Understanding of Online Comics
by: Ryan, Yuriel, et al.
Published: (2025)

Examining the Influence of Political Bias on Large Language Model Performance in Stance Classification
by: Ng, Lynnette Hui Xian, et al.
Published: (2024)

SuperWriter: Reflection-Driven Long-Form Generation with Large Language Models
by: Wu, Yuhao, et al.
Published: (2025)

OR-Bench: An Over-Refusal Benchmark for Large Language Models
by: Cui, Justin, et al.
Published: (2024)

InstructAV: Instruction Fine-tuning Large Language Models for Authorship Verification
by: Hu, Yujia, et al.
Published: (2024)

HateXScore: A Metric Suite for Evaluating Reasoning Quality in Hate Speech Explanations
by: Hu, Yujia, et al.
Published: (2026)

SteerRM: Debiasing Reward Models via Sparse Autoencoders
by: Sun, Mengyuan, et al.
Published: (2026)

SusGen-GPT: A Data-Centric LLM for Financial NLP and Sustainability Report Generation
by: Wu, Qilong, et al.
Published: (2024)

Large Language Models for Automated Open-domain Scientific Hypotheses Discovery
by: Yang, Zonglin, et al.
Published: (2023)

SAFER: Probing Safety in Reward Models with Sparse Autoencoder
by: Shi, Wei, et al.
Published: (2025)

Large Language Models for Few-Shot Named Entity Recognition
by: Zhao, Yufei, et al.
Published: (2018)

Language Models as Inductive Reasoners
by: Yang, Zonglin, et al.
Published: (2022)

XAI meets LLMs: A Survey of the Relation between Explainable AI and Large Language Models
by: Cambria, Erik, et al.
Published: (2024)

Toxicity Red-Teaming: Benchmarking LLM Safety in Singapore's Low-Resource Languages
by: Hu, Yujia, et al.
Published: (2025)

Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders
by: Deng, Boyi, et al.
Published: (2025)

Enhancing Language Models for Robust Greenwashing Detection
by: Braun, Neil Heinrich, et al.
Published: (2026)

ToxiCloakCN: Evaluating Robustness of Offensive Language Detection in Chinese with Cloaking Perturbations
by: Xiao, Yunze, et al.
Published: (2024)

Bridging Modalities: Enhancing Cross-Modality Hate Speech Detection with Few-Shot In-Context Learning
by: Hee, Ming Shan, et al.
Published: (2024)

Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
by: Jain, Neel, et al.
Published: (2024)

Modularized Networks for Few-shot Hateful Meme Detection
by: Cao, Rui, et al.
Published: (2024)

Towards Understanding the Robustness of Sparse Autoencoders
by: Saiyed, Ahson, et al.
Published: (2026)

ProtSAE: Disentangling and Interpreting Protein Language Models via Semantically-Guided Sparse Autoencoders
by: Liu, Xiangyu, et al.
Published: (2025)

SEAHateCheck: Functional Tests for Detecting Hate Speech in Low-Resource Languages of Southeast Asia
by: Ng, Ri Chi, et al.
Published: (2026)