# Does Reasoning Introduce Bias? A Study of Social Bias Evaluation and Mitigation in LLM Reasoning Source: [Does Reasoning Introduce Bias? A Study of Social Bias Evaluation and Mitigation in LLM Reasoning]() Collection index: [llms.txt]() Data updated: 2025-10-26T05:41:29.127Z Authors: Xuyang Wu, Jinming Nian, Ting-Ruen Wei, Zhiqiang Tao, Hsin-Tai Wu, Yi Fang Published: 2025-02-21T10:16:07.000Z ## Abstract Recent advances in large language models (LLMs) have enabled automatic generation of chain-of-thought (CoT) reasoning, leading to strong performance on tasks such as math and code. However, when reasoning steps reflect social stereotypes (e.g., those related to gender, race or age), they can reinforce harmful associations and lead to misleading conclusions. We present the first systematic evaluation of social bias within LLM-generated reasoning, focusing on reasoning language models (e.g., DeepSeek-R1, OpenAI o1) that natively produce reasoning chains as part of their answers. Using the BBQ dataset, we analyze both prediction accuracy and reasoning bias across a broad spectrum of models, including instruction-tuned and CoT-augmented variants of DeepSeek-R1 (8B/32B), ChatGPT, and other open-source LLMs. We quantify how biased reasoning steps correlate with incorrect predictions and often lead to stereotype expression. To mitigate reasoning-induced bias, we propose Answer Distribution as Bias Proxy (ADBP), a lightweight mitigation method that detects bias by tracking how model predictions change across incremental reasoning steps. ADBP outperforms Stereotype-free Reasoning Pattern (SfRP) baseline in most cases, mitigating bias and improving the accuracy of LLM outputs. Evaluation and mitigation code is available at https://github.com/elviswxy/LLM\_reasoning\_bias. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Fairness Evaluation on Fairness Evaluation Suite (BBQ, CrP, GMO, SSt, WnQ)](): Fairness evaluation; Fairness Evaluation Suite BBQ, CrP, GMO, SSt, WnQ; BBQ Score: 97.99. - [Question Answering on Fairness Evaluation Suite BBQ, CrP, GMO, SSt, WnQ](): Question Answering; Fairness Evaluation Suite BBQ, CrP, GMO, SSt, WnQ; BBQ Accuracy: 97.99. ## Original sources and implementation [Original paper]() [PDF]()