Bias Benchmark

Bias benchmarks are tools used to evaluate and mitigate biases in large language models (LLMs), focusing on identifying and quantifying societal prejudices like gender, racial, and religious biases, as well as biases related to sexual orientation, nationality, and other social categories. Current research emphasizes developing more holistic benchmarks that account for contextual factors, intertwined biases across different domains, and the limitations of using LLMs themselves as evaluators due to their inherent cognitive biases. This work is crucial for ensuring fairness and ethical considerations in the development and deployment of LLMs, impacting both the scientific understanding of AI bias and the creation of more equitable AI systems.

Papers

November 6, 2024

The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models
Anaelia Ovalle, Krunoslav Lehman Pavasovic, Louis Martin, Luke Zettlemoyer, Eric Michael Smith, Adina Williams, Levent Sagun
Language Model Gender Bias Bias Metric Bias Benchmark FRUIT Produce SMOOTHIE

November 1, 2024

Benchmarking Bias in Large Language Models during Role-Playing
Xinyue Li, Zhenpeng Chen, Jie M. Zhang, Yiling Lou, Tianlin Li, Weisong Sun, Yang Liu, Xuanzhe Liu
Social Bias Role Play Bias Benchmark Challenging Role Type

October 14, 2024

Assessing Bias in Metric Models for LLM Open-Ended Generation Bias Benchmarks
Nathaniel Demchak, Xin Guan, Zekun Wu, Ziyi Xu, Adriano Koshiyama, Emre Kazim
New Benchmark Social Bias Stereotype Content Bias Evaluation Inherent Bias Demographic Attribute Bias Benchmark LLM Bias

October 9, 2024

VHELM: A Holistic Evaluation of Vision Language Models
Tony Lee, Haoqin Tu, Chi Heem Wong, Wenhao Zheng, Yiyang Zhou, Yifan Mai, Josselin Somerville Roberts, Michihiro Yasunaga, Huaxiu Yao, Cihang Xie, Percy Liang
Vision Language Model Evaluation Metric Visual Perception Holistic Evaluation Bias Benchmark

September 25, 2024

Mitigating the Bias of Large Language Model Evaluation
Hongli Zhou, Hui Huang, Yunfei Long, Bing Xu, Conghui Zhu, Hailong Cao, Muyun Yang, Tiejun Zhao
Large Language Model Absolute Stance Bias Contrastive Training Bias Benchmark Fine Tuned Judge Model

September 17, 2024

SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration
Xin Guan, Nathaniel Demchak, Saloni Gupta, Ze Wang, Ediz Ertekin Jr., Adriano Koshiyama, Emre Kazim, Zekun Wu
Language Model Contextual Biasing Role Playing Bias Benchmark Bias Calibration

August 27, 2024

Intertwined Biases Across Social Media Spheres: Unpacking Correlations in Media Bias Dimensions
Yifan Liu, Yike Li, Dong Wang
Total Correlation Medium Bias Bias Benchmark Correlation Bias

July 15, 2024

Foundational Autoraters: Taming Large Language Models for Better Automatic Evaluation
Tu Vu, Kalpesh Krishna, Salaheddin Alzubi, Chris Tar, Manaal Faruqui, Yun-Hsuan Sung
Large Language Model Evaluation Benchmark Automatic Evaluation LLM Evaluator Bias Benchmark

July 3, 2024

Social Bias in Large Language Models For Bangla: An Empirical Study on Gender and Religious Bias
Jayanta Sadhu, Maneesha Rani Saha, Rifat Shahriyar
Social Bias Bangla Text Gender Inclusive Text Bias Evaluation BIAS Detection Bias Benchmark Bangla Natural Language

June 4, 2024

Analyzing Social Biases in Japanese Large Language Models
Hitomi Yanaka, Namgi Han, Ryoma Kumon, Jie Lu, Masashi Takeshita, Ryo Sekizawa, Taisei Kato, Hiromi Arai
Question Answering Social Bias Bias Benchmark Japanese Large Language Model

May 24, 2024

GPT is Not an Annotator: The Necessity of Human Annotation in Fairness Benchmark Construction
Virginia K. Felkner, Jennifer A. Thompson, Jonathan May
Human Annotation Social Bias GPT Neo Community Need Individual Annotator Bias Benchmark Fairness Benchmark

February 22, 2024

COBIAS: Contextual Reliability in Bias Assessment
Priyanshul Govil, Hemang Jain, Vamshi Krishna Bonagiri, Aman Chadha, Ponnurangam Kumaraguru, Manas Gaur, Sanorita Dey
Bias Evaluation Contextual Biasing Model Debiasing Bias Benchmark Contextual Reliability

February 20, 2024

Bias in Language Models: Beyond Trick Tests and Toward RUTEd Evaluation
Kristian Lum, Jacy Reese Anthis, Chirag Nagpal, Alexander D'Amour
Language Model Absolute Stance Bias Bias Evaluation Unconventional Rabbit Hat Trick Bias Benchmark

November 21, 2023

Benchmarking bias: Expanding clinical AI model card to incorporate bias reporting of social and non-social factors
Carolina A. M. Heming, Mohamed Abdalla, Monish Ahluwalia, Linglin Zhang, Hari Trivedi, MinJae Woo, Benjamin Fine, Judy Wawira Gichoya, Leo Anthony Celi, Laleh Seyyed-Kalantari
Key Factor Topic Bias AI Model Clinical Artificial Intelligence Bias Benchmark Artificial Intelligence Bias Source Bias Social Factor

September 29, 2023

Benchmarking Cognitive Biases in Large Language Models as Evaluators
Ryan Koo, Minhwa Lee, Vipul Raheja, Jong Inn Park, Zae Myung Kim, Dongyeop Kang
Human Evaluation Bias Benchmark Text Quality

September 11, 2023

Challenges in Annotating Datasets to Quantify Bias in Under-represented Society
Vithya Yogarajan, Gillian Dobbie, Timothy Pistotti, Joshua Bensemann, Kobe Knowles
Technical Challenge Absolute Stance Bias Annotated Dataset Benchmark Datasets Underrepresented Group Potential Bias Bias Benchmark Quantifying Bias Dataset Annotation

July 31, 2023

KoBBQ: Korean Bias Benchmark for Question Answering
Jiho Jin, Jiseon Kim, Nayeon Lee, Haneul Yoo, Alice Oh, Hwaran Lee
Language Model Question Answering Bias Benchmark Social Bias Benchmark

July 15, 2023

Bound by the Bounty: Collaboratively Shaping Evaluation Processes for Queer AI Harms
Organizers of QueerInAI, Nathan Dennler, Anaelia Ovalle, Ashwin Singh, Luca Soldaini, Arjun Subramonian, Huy Tu, William Agnew, Avijit Ghosh, Kyra Yee, Irene Font Peradejordi, Zeerak Talat, Mayra Russo, Jess de Jesus de Pinho Pinhal
Artificial Intelligence Bias Evaluation Bias Benchmark Participatory Approach AI Harm

June 28, 2023

CBBQ: A Chinese Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models
Yufei Huang, Deyi Xiong
Large Language Model Human Ai Collaboration Bias Evaluation Societal Bias Bias Benchmark Chinese Large Language Model

June 26, 2023

WinoQueer: A Community-in-the-Loop Benchmark for Anti-LGBTQ+ Bias in Large Language Models
Virginia K. Felkner, Ho-Chun Herbert Chang, Eugene Jang, Jonathan May
Large Language Model Bias Benchmark Encode Bias Anti LGBTQ+

Bias Benchmark

Papers

The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models

Benchmarking Bias in Large Language Models during Role-Playing

Assessing Bias in Metric Models for LLM Open-Ended Generation Bias Benchmarks

VHELM: A Holistic Evaluation of Vision Language Models

Mitigating the Bias of Large Language Model Evaluation

SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration

Intertwined Biases Across Social Media Spheres: Unpacking Correlations in Media Bias Dimensions

Foundational Autoraters: Taming Large Language Models for Better Automatic Evaluation

Social Bias in Large Language Models For Bangla: An Empirical Study on Gender and Religious Bias

Analyzing Social Biases in Japanese Large Language Models

GPT is Not an Annotator: The Necessity of Human Annotation in Fairness Benchmark Construction

COBIAS: Contextual Reliability in Bias Assessment

Bias in Language Models: Beyond Trick Tests and Toward RUTEd Evaluation

Benchmarking bias: Expanding clinical AI model card to incorporate bias reporting of social and non-social factors

Benchmarking Cognitive Biases in Large Language Models as Evaluators

Challenges in Annotating Datasets to Quantify Bias in Under-represented Society

KoBBQ: Korean Bias Benchmark for Question Answering

Bound by the Bounty: Collaboratively Shaping Evaluation Processes for Queer AI Harms

CBBQ: A Chinese Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models

WinoQueer: A Community-in-the-Loop Benchmark for Anti-LGBTQ+ Bias in Large Language Models