Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GIEBench: Towards Holistic Evaluation of Group Identity-based Empathy for Large Language Models

Leyan Wang, Yonggang Jin, Tianhao Shen, Tianyu Zheng, Xinrun Du, Chenchen Zhang, Wenhao Huang, Jiaheng Liu, Shi Wang, Ge Zhang, Liuyu Xiang, Zhaofeng He

TL;DR

GIEBench is a comprehensive benchmark that includes 11 identity dimensions, covering 97 group identities with a total of 999 single-choice questions related to specific group identities, designed to evaluate the empathy of LLMs when presented with specific group identities such as gender, age, occupation, and race.

Abstract

As large language models (LLMs) continue to develop and gain widespread application, the ability of LLMs to exhibit empathy towards diverse group identities and understand their perspectives is increasingly recognized as critical. Most existing benchmarks for empathy evaluation of LLMs focus primarily on universal human emotions, such as sadness and pain, often overlooking the context of individuals' group identities. To address this gap, we introduce GIEBench, a comprehensive benchmark that includes 11 identity dimensions, covering 97 group identities with a total of 999 single-choice questions related to specific group identities. GIEBench is designed to evaluate the empathy of LLMs when presented with specific group identities such as gender, age, occupation, and race, emphasizing their ability to respond from the standpoint of the identified group. This supports the ongoing development of empathetic LLM applications tailored to users with different identities. Our evaluation of 23 LLMs revealed that while these LLMs understand different identity standpoints, they fail to consistently exhibit equal empathy across these identities without explicit instructions to adopt those perspectives. This highlights the need for improved alignment of LLMs with diverse values to better accommodate the multifaceted nature of human identities. Our datasets are available at https://github.com/GIEBench/GIEBench.

GIEBench: Towards Holistic Evaluation of Group Identity-based Empathy for Large Language Models

TL;DR

GIEBench is a comprehensive benchmark that includes 11 identity dimensions, covering 97 group identities with a total of 999 single-choice questions related to specific group identities, designed to evaluate the empathy of LLMs when presented with specific group identities such as gender, age, occupation, and race.

Abstract

As large language models (LLMs) continue to develop and gain widespread application, the ability of LLMs to exhibit empathy towards diverse group identities and understand their perspectives is increasingly recognized as critical. Most existing benchmarks for empathy evaluation of LLMs focus primarily on universal human emotions, such as sadness and pain, often overlooking the context of individuals' group identities. To address this gap, we introduce GIEBench, a comprehensive benchmark that includes 11 identity dimensions, covering 97 group identities with a total of 999 single-choice questions related to specific group identities. GIEBench is designed to evaluate the empathy of LLMs when presented with specific group identities such as gender, age, occupation, and race, emphasizing their ability to respond from the standpoint of the identified group. This supports the ongoing development of empathetic LLM applications tailored to users with different identities. Our evaluation of 23 LLMs revealed that while these LLMs understand different identity standpoints, they fail to consistently exhibit equal empathy across these identities without explicit instructions to adopt those perspectives. This highlights the need for improved alignment of LLMs with diverse values to better accommodate the multifaceted nature of human identities. Our datasets are available at https://github.com/GIEBench/GIEBench.
Paper Structure (26 sections, 3 figures, 8 tables)

This paper contains 26 sections, 3 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Empathy Evaluation of LLMs
  4. Value Pluralism of LLMs
  5. GIEBench
  6. Plural Controversial Topics Generation
  7. Internet Sourcing
  8. GPT-4 Based Synthetic Topic Generation
  9. Human Annotation
  10. Prompt Construction and Pipeline
  11. Data Statistics
  12. Results and Analysis
  13. Evaluating Plurality in LLMs
  14. Experiment Settings
  15. Main Results
  16. ...and 11 more sections

Figures (3)

  • Figure 1: The proportion of the eleven identity dimensions in GIEBench. The categories of Gender and Occupation have the smallest proportions, accounting for 6.71% and 6%, respectively. The proportions of the remaining categories are all around 10%. A broad range of categories facilitates our evaluation of LLMs’ performance across various identity standpoints.
  • Figure 2: The process of constructing GIEBench. Initially, a collection of controversial topics is developed using web resources, manual selection, and GPT-4, each corresponding to a specific identity. Subsequently, we annotate attitude labels from the perspectives of these identities. We also utilize GPT-4 to generate four responses for each topic, ensuring that only one response aligns with the identity's stance. Finally, using the established identities, topics, and responses, we design three types of prompts to LLMs in selecting the most appropriate response. In the COT-Prompt, a Chain of Thought (COT) is provided along with identity information. In the ID-Prompt, only the identity is disclosed, while the Raw-Prompt includes no additional information.
  • Figure 3: The figure illustrates the accuracy of six different series of Large Language Models (LLMs) on our dataset based on COT-Prompt. Overall, GPT-4-turbo performs better, which, to some extent, reflects its superior alignment across various identity positions.