Heterogeneous Value Alignment Evaluation for Large Language Models

TL;DR

This paper introduces a novel evaluation system for assessing how well large language models align with diverse human values, emphasizing the importance of transferring heterogeneous values in practical applications.

Contribution

It proposes the Heterogeneous Value Alignment Evaluation (HVAE) system that incorporates social value orientation to measure LLMs' ability to pursue and align with different values.

Findings

LLMs tend to favor neutral over personal values.

The HVAE system effectively measures value rationality in LLMs.

Insights into LLMs' value alignment within heterogeneous systems.

Abstract

The emergent capabilities of Large Language Models (LLMs) have made it crucial to align their values with those of humans. However, current methodologies typically attempt to assign value as an attribute to LLMs, yet lack attention to the ability to pursue value and the importance of transferring heterogeneous values in specific practical applications. In this paper, we propose a Heterogeneous Value Alignment Evaluation (HVAE) system, designed to assess the success of aligning LLMs with heterogeneous values. Specifically, our approach first brings the Social Value Orientation (SVO) framework from social psychology, which corresponds to how much weight a person attaches to the welfare of others in relation to their own. We then assign the LLMs with different social values and measure whether their behaviors align with the inducing values. We conduct evaluations with new auto-metric \textit{value rationality} to represent the ability of LLMs to align with specific values. Evaluating the value rationality of five mainstream LLMs, we discern a propensity in LLMs towards neutral values over pronounced personal values. By examining the behavior of these LLMs, we contribute to a deeper insight into the value alignment of LLMs within a heterogeneous value system.