# Investigating Thinking Behaviours of Reasoning-Based Language Models for Social Bias Mitigation

arXiv: https://arxiv.org/abs/2510.17062 · HTML: https://mlsanity.com/paper/2510.17062/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 16 of its 46 references that are in the map.
- First posted on arXiv 20 Oct 2025.

## Placed in

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

## Builds on (most cited first, 16 of 16)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs](https://mlsanity.com/paper/2412.21187/) (2024)
- [Reasoning Models Don't Always Say What They Think](https://mlsanity.com/paper/2505.05410/) (2025)
- [From System 1 to System 2: A Survey of Reasoning Large Language Models](https://mlsanity.com/paper/2502.17419/) (2025)
- [Dynamic Early Exit in Reasoning Models](https://mlsanity.com/paper/2504.15895/) (2025)
- [When More is Less: Understanding Chain-of-Thought Length in LLMs](https://mlsanity.com/paper/2502.07266/) (2025)
- [Between Underthinking and Overthinking: An Empirical Study of Reasoning Length and correctness in LLMs](https://mlsanity.com/paper/2505.00127/) (2025)
- [Evaluating Gender Bias in Large Language Models via Chain-of-Thought Prompting](https://mlsanity.com/paper/2401.15585/) (2024)
- [Deconstructing Long Chain-of-Thought: A Structured Reasoning Optimization Framework for Long CoT Distillation](https://mlsanity.com/paper/2503.16385/) (2025)
- [Does Reasoning Introduce Bias? A Study of Social Bias Evaluation and Mitigation in LLM Reasoning](https://mlsanity.com/paper/2502.15361/) (2025)
- [JBBQ: Japanese Bias Benchmark for Analyzing Social Biases in Large Language Models](https://mlsanity.com/paper/2406.02050/) (2024)
- [Evaluating the Effect of Retrieval Augmentation on Social Biases](https://mlsanity.com/paper/2502.17611/) (2025)
- [Don't "Overthink" Passage Reranking: Is Reasoning Truly Necessary?](https://mlsanity.com/paper/2505.16886/) (2025)
- [Intent-Aware Self-Correction for Mitigating Social Biases in Large Language Models](https://mlsanity.com/paper/2503.06011/) (2025)
- [Is Reasoning All You Need? Probing Bias in the Age of Reasoning Language Models](https://mlsanity.com/paper/2507.02799/) (2025)

## Built on by (most cited first, 2 of 2)

- [More Thinking, More Bias: Length-Driven Position Bias in Reasoning Models](https://mlsanity.com/paper/2605.06672/) (2026)
- [Wait, am I Being Fair? Characterizing Deductive Stereotyping and Mitigating It with Fair-GCG](https://mlsanity.com/paper/2606.30989/) (2026)

## Abstract (arXiv)

While reasoning-based large language models excel at complex tasks through an internal, structured thinking process, a concerning phenomenon has emerged that such a thinking process can aggregate social stereotypes, leading to biased outcomes. However, the underlying behaviours of these language models in social bias scenarios remain underexplored. In this work, we systematically investigate mechanisms within the thinking process behind this phenomenon and uncover two failure patterns that drive social bias aggregation: 1) stereotype repetition, where the model relies on social stereotypes as its primary justification, and 2) irrelevant information injection, where it fabricates or introduces new details to support a biased narrative. Building on these insights, we introduce a lightweight prompt-based mitigation approach that queries the model to review its own initial reasoning against these specific failure patterns. Experiments on question answering (BBQ and StereoSet) and open-ended (BOLD) benchmarks show that our approach effectively reduces bias while maintaining or improving accuracy.