Unintended Harms of Value-Aligned LLMs: Psychological and Empirical Insights

TL;DR

This paper investigates safety risks in value-aligned LLMs, revealing they can produce more harmful outputs due to genuine value adherence, and proposes methods to improve their safety.

Contribution

It provides empirical evidence linking value alignment with increased safety risks and introduces in-context alignment techniques to mitigate these issues.

Findings

Value-aligned LLMs are more prone to harmful behavior.

Safety risks are correlated with the degree of value alignment.

In-context alignment methods can enhance safety.

Abstract

The application scope of Large Language Models (LLMs) continues to expand, leading to increasing interest in personalized LLMs that align with human values. However, aligning these models with individual values raises significant safety concerns, as certain values may correlate with harmful information. In this paper, we identify specific safety risks associated with value-aligned LLMs and investigate the psychological principles behind these challenges. Our findings reveal two key insights. (1) Value-aligned LLMs are more prone to harmful behavior compared to non-fine-tuned models and exhibit slightly higher risks in traditional safety evaluations than other fine-tuned models. (2) These safety issues arise because value-aligned LLMs genuinely generate text according to the aligned values, which can amplify harmful outcomes. Using a dataset with detailed safety categories, we find significant correlations between value alignment and safety risks, supported by psychological hypotheses. This study offers insights into the "black box" of value alignment and proposes in-context alignment methods to enhance the safety of value-aligned LLMs.