Improving Large Language Model Safety with Contrastive Representation Learning

ACL ARR 2025 May Submission4815 Authors

Abstract:

Large Language Models (LLMs) are powerful tools with profound societal impacts,
yet their ability to generate responses to diverse and uncontrolled inputs
leaves them vulnerable to adversarial attacks. While existing defenses often
struggle to generalize across varying attack types, recent advancements in
representation engineering offer promising alternatives. In this work, we
propose a defense framework that formulates model defense as a contrastive
representation learning (CRL) problem. Our method finetunes a model using
a triplet-based loss combined with adversarial hard negative mining to encourage
separation between benign and harmful representations. Our experimental results
across multiple models demonstrate that our approach outperforms prior
representation engineering-based defenses, improving robustness against both
input-level and embedding-space attacks without compromising standard
performance.

Paper Type: Long

Research Area: Language Modeling

Research Area Keywords: red teaming,safety and alignment,security and privacy,robustness

Contribution Types: NLP engineering experiment, Publicly available software and/or pre-trained models

Languages Studied: English

Submission Number: 4815