With a Grain of SALT: Are LLMs Fair Across Social Dimensions?
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917926901121024 |
|---|---|
| author | Arif, Samee Khan, Zohaib Kaleem, Maaidah Rashid, Suhaib Raza, Agha Ali Athar, Awais |
| author_facet | Arif, Samee Khan, Zohaib Kaleem, Maaidah Rashid, Suhaib Raza, Agha Ali Athar, Awais |
| contents | This paper presents a systematic analysis of biases in open-source Large Language Models (LLMs), across gender, religion, and race. Our study evaluates bias in smaller-scale Llama and Gemma models using the SALT ($\textbf{S}$ocial $\textbf{A}$ppropriateness in $\textbf{L}$LM-Generated $\textbf{T}$ext) dataset, which incorporates five distinct bias triggers: General Debate, Positioned Debate, Career Advice, Problem Solving, and CV Generation. To quantify bias, we measure win rates in General Debate and the assignment of negative roles in Positioned Debate. For real-world use cases, such as Career Advice, Problem Solving, and CV Generation, we anonymize the outputs to remove explicit demographic identifiers and use DeepSeek-R1 as an automated evaluator. We also address inherent biases in LLM-based evaluation, including evaluation bias, positional bias, and length bias, and validate our results through human evaluations. Our findings reveal consistent polarization across models, with certain demographic groups receiving systematically favorable or unfavorable treatment. By introducing SALT, we provide a comprehensive benchmark for bias analysis and underscore the need for robust bias mitigation strategies in the development of equitable AI systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_12499 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | With a Grain of SALT: Are LLMs Fair Across Social Dimensions? Arif, Samee Khan, Zohaib Kaleem, Maaidah Rashid, Suhaib Raza, Agha Ali Athar, Awais Computation and Language This paper presents a systematic analysis of biases in open-source Large Language Models (LLMs), across gender, religion, and race. Our study evaluates bias in smaller-scale Llama and Gemma models using the SALT ($\textbf{S}$ocial $\textbf{A}$ppropriateness in $\textbf{L}$LM-Generated $\textbf{T}$ext) dataset, which incorporates five distinct bias triggers: General Debate, Positioned Debate, Career Advice, Problem Solving, and CV Generation. To quantify bias, we measure win rates in General Debate and the assignment of negative roles in Positioned Debate. For real-world use cases, such as Career Advice, Problem Solving, and CV Generation, we anonymize the outputs to remove explicit demographic identifiers and use DeepSeek-R1 as an automated evaluator. We also address inherent biases in LLM-based evaluation, including evaluation bias, positional bias, and length bias, and validate our results through human evaluations. Our findings reveal consistent polarization across models, with certain demographic groups receiving systematically favorable or unfavorable treatment. By introducing SALT, we provide a comprehensive benchmark for bias analysis and underscore the need for robust bias mitigation strategies in the development of equitable AI systems. |
| title | With a Grain of SALT: Are LLMs Fair Across Social Dimensions? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2410.12499 |