# SODAPOP: Open-Ended Discovery of Social Biases in Social Commonsense Reasoning Models

arXiv: https://arxiv.org/abs/2210.07269 · HTML: https://mlsanity.com/paper/2210.07269/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 38 of its 55 references that are in the map.
- First posted on arXiv 13 Oct 2022.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 30 of 38)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge](https://mlsanity.com/paper/1811.00937/) (2018)
- [Language (Technology) is Power: A Critical Survey of "Bias" in NLP](https://mlsanity.com/paper/2005.14050/) (2020)
- [StereoSet: Measuring stereotypical bias in pretrained language models](https://mlsanity.com/paper/2004.09456/) (2020)
- [Semantics derived automatically from language corpora contain human-like biases](https://mlsanity.com/paper/1608.07187/) (2016)
- [How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings](https://mlsanity.com/paper/1909.00512/) (2019)
- [CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models](https://mlsanity.com/paper/2010.00133/) (2020)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection](https://mlsanity.com/paper/2004.07667/) (2020)
- [COMET: Commonsense Transformers for Automatic Knowledge Graph Construction](https://mlsanity.com/paper/1906.05317/) (2019)
- [Gender Bias in Coreference Resolution](https://mlsanity.com/paper/1804.09301/) (2018)
- [A Corpus and Evaluation Framework for Deeper Understanding of Commonsense Stories](https://mlsanity.com/paper/1604.01696/) (2016)
- [On Measuring Social Biases in Sentence Encoders](https://mlsanity.com/paper/1903.10561/) (2019)
- [Social Bias Frames: Reasoning about Social and Power Implications of Language](https://mlsanity.com/paper/1911.03891/) (2019)
- [Measuring Bias in Contextualized Word Representations](https://mlsanity.com/paper/1906.07337/) (2019)
- [Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them](https://mlsanity.com/paper/1903.03862/) (2019)
- [Social Chemistry 101: Learning to Reason about Social and Moral Norms](https://mlsanity.com/paper/2011.00620/) (2020)
- [Measuring and Reducing Gendered Correlations in Pre-trained Models](https://mlsanity.com/paper/2010.06032/) (2020)
- [Gender Bias in Contextualized Word Embeddings](https://mlsanity.com/paper/1904.03310/) (2019)
- [Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology](https://mlsanity.com/paper/1906.04571/) (2019)
- [Towards Debiasing Sentence Representations](https://mlsanity.com/paper/2007.08100/) (2020)
- [Detecting Emergent Intersectional Biases: Contextualized Word Embeddings Contain a Distribution of Human-like Biases](https://mlsanity.com/paper/2006.03955/) (2020)
- [UnQovering Stereotyping Biases via Underspecified Questions](https://mlsanity.com/paper/2010.02428/) (2020)
- [Towards Controllable Biases in Language Generation](https://mlsanity.com/paper/2005.00268/) (2020)
- [It's All in the Name: Mitigating Gender Bias with Name-Based Counterfactual Data Substitution](https://mlsanity.com/paper/1909.00871/) (2019)
- [An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-trained Language Models](https://mlsanity.com/paper/2110.08527/) (2021)
- [Quantifying Social Biases in NLP: A Generalization and Empirical Comparison of Extrinsic Fairness Metrics](https://mlsanity.com/paper/2106.14574/) (2021)
- ["You are grounded!": Latent Name Artifacts in Pre-trained Language Models](https://mlsanity.com/paper/2004.03012/) (2020)
- [Low Frequency Names Exhibit Bias and Overfitting in Contextualizing Language Models](https://mlsanity.com/paper/2110.00672/) (2021)

## Built on by (most cited first, 22 of 22)

- [Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models](https://mlsanity.com/paper/2305.18189/) (2023)
- [Towards Measuring and Modeling "Culture" in LLMs: A Survey](https://mlsanity.com/paper/2403.15412/) (2024)
- [Having Beer after Prayer? Measuring Cultural Bias in Large Language Models](https://mlsanity.com/paper/2305.14456/) (2023)
- [Fairness in Large Language Models: A Taxonomic Survey](https://mlsanity.com/paper/2404.01349/) (2024)
- [Social Bias Evaluation for Large Language Models Requires Prompt Variations](https://mlsanity.com/paper/2407.03129/) (2024)
- [Sociodemographic Bias in Language Models: A Survey and Forward Path](https://mlsanity.com/paper/2306.08158/) (2023)
- [Culturally Aware and Adapted NLP: A Taxonomy and a Survey of the State of the Art](https://mlsanity.com/paper/2406.03930/) (2024)
- [Do Large Language Models Discriminate in Hiring Decisions on the Basis of Race, Ethnicity, and Gender?](https://mlsanity.com/paper/2406.10486/) (2024)
- [Stop! In the Name of Flaws: Disentangling Personal Names and Sociodemographic Attributes in NLP](https://mlsanity.com/paper/2405.17159/) (2024)
- [Cultural Conditioning or Placebo? On the Effectiveness of Socio-Demographic Prompting](https://mlsanity.com/paper/2406.11661/) (2024)
- [CALM : A Multi-task Benchmark for Comprehensive Assessment of Language Model Bias](https://mlsanity.com/paper/2308.12539/) (2023)
- [Nichelle and Nancy: The Influence of Demographic Attributes and Tokenization Length on First Name Biases](https://mlsanity.com/paper/2305.16577/) (2023)
- [Fairness Definitions in Language Models Explained](https://mlsanity.com/paper/2407.18454/) (2024)
- [mCSQA: Multilingual Commonsense Reasoning Dataset with Unified Creation Strategy by Language Models and Humans](https://mlsanity.com/paper/2406.04215/) (2024)
- [CLIMB: A Benchmark of Clinical Bias in Large Language Models](https://mlsanity.com/paper/2407.05250/) (2024)
- [On the Mutual Influence of Gender and Occupation in LLM Representations](https://mlsanity.com/paper/2503.06792/) (2025)
- [With a Grain of SALT: Are LLMs Fair Across Social Dimensions?](https://mlsanity.com/paper/2410.12499/) (2024)
- [When Stereotypes GTG: The Impact of Predictive Text Suggestions on Gender Bias in Human-AI Co-Writing](https://mlsanity.com/paper/2409.20390/) (2024)
- [Large Language Models Struggle to Describe the Haystack without Human Help: Human-in-the-loop Evaluation of Topic Models](https://mlsanity.com/paper/2502.14748/) (2025)
- [Large Language Models Are Effective Human Annotation Assistants, But Not Good Independent Annotators](https://mlsanity.com/paper/2503.06778/) (2025)
- [On the Influence of Gender and Race in Romantic Relationship Prediction from Large Language Models](https://mlsanity.com/paper/2410.03996/) (2024)
- [Developing A Framework to Support Human Evaluation of Bias in Generated Free Response Text](https://mlsanity.com/paper/2505.03053/) (2025)

## Abstract (arXiv)

A common limitation of diagnostic tests for detecting social biases in NLP models is that they may only detect stereotypic associations that are pre-specified by the designer of the test. Since enumerating all possible problematic associations is infeasible, it is likely these tests fail to detect biases that are present in a model but not pre-specified by the designer. To address this limitation, we propose SODAPOP (SOcial bias Discovery from Answers about PeOPle) in social commonsense question-answering. Our pipeline generates modified instances from the Social IQa dataset (Sap et al., 2019) by (1) substituting names associated with different demographic groups, and (2) generating many distractor answers from a masked language model. By using a social commonsense model to score the generated distractors, we are able to uncover the model's stereotypic associations between demographic groups and an open set of words. We also test SODAPOP on debiased models and show the limitations of multiple state-of-the-art debiasing algorithms.