mlsanityData to 14 Oct 2026

Does Reasoning Introduce Bias? A Study of Social Bias Evaluation and Mitigation in LLM Reasoning

arXiv 2502.15361 ↗First posted 21 Feb 2025cs.CL cs.AIcited by 21

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

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.

Key facts. Cited by 21 papers in the map (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 13 of its 39 references that are in the map. First posted on arXiv 21 Feb 2025.

Builds on

13 of 39 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024Evaluating Gender Bias in Large Language Models via Chain-of-Thought Prompting28 Jan

Other references · 12

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  3. 2024Qwen2 Technical Reportcited by 2,120
  4. 2024Monte Carlo Tree Search Boosts Reasoning via Iterative Preference Learningcited by 241
  5. 2025Thoughts Are All Over the Place: On the Underthinking of o1-Like LLMscited by 163
  6. 2024Marco-o1: Towards Open Reasoning Models for Open-Ended Solutionscited by 144
  7. 2024Decoding Biases: Automated Methods and LLM Judges for Gender Bias Detection in Language Modelscited by 33
  8. 2023Prompting Frameworks for Large Language Models: A Surveycited by 29
  9. 2023Mathematical Language Models: A Surveycited by 24
  10. 2024Evaluating Fairness in Large Vision-Language Models Across Diverse Demographic Attributes and Promptscited by 15
  11. 2025A Survey on Feedback-based Multi-step Reasoning for Large Language Models on Mathematicscited by 11
  12. 2024Evaluating and Mitigating Social Bias for Large Language Models in Open-ended Settingscited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 8; 2026: 13.

21 papers.

Sort:

By area:

  1. 2026Self-Preference Bias in Rubric-Based Evaluation of Large Language Modelscited by 11
  2. 2025Schoenfeld's Anatomy of Mathematical Reasoning by Language Modelscited by 8
  3. 2026Reward Modeling for Reinforcement Learning-Based LLM Reasoning: Design, Challenges, and Evaluationcited by 7
  4. 2025Is Reasoning All You Need? Probing Bias in the Age of Reasoning Language Modelscited by 4
  5. 2025Cognitive Chain-of-Thought (CoCoT): Structured Multimodal Reasoning about Social Situationscited by 3
  6. 2025Test-Time Reasoners Are Strategic Multiple-Choice Test-Takerscited by 3
  7. 2025BiasCause: Evaluate Socially Biased Causal Reasoning of Large Language Modelscited by 2
  8. 2025Think Again! The Effect of Test-Time Compute on Preferences, Opinions, and Beliefs of Large Language Modelscited by 2
  9. 2025Investigating Thinking Behaviours of Reasoning-Based Language Models for Social Bias Mitigationcited by 2
  10. 2026Inference-Time Reasoning Selectively Reduces Implicit Social Bias in Large Language Modelscited by 2
  11. 2026Calibrated? Not for Everyone: How Sexual Orientation and Religious Markers Distort LLM Accuracy and Confidence in Medical QAcited by 2
  12. 2026Taming Actor-Observer Asymmetry in Agents via Dialectical Alignmentcited by 2
9 more
  1. 2025Evaluating Implicit Biases in LLM Reasoning through Logic Grid Puzzlescited by 1
  2. 2026QuantiBias: Benchmarking Quantization-Induced Bias in LLMscited by 1
  3. 2026The Periodic Table of LLM Reasoning: A Structured Survey of Reasoning Paradigms, Methods, and Failure Modescited by 0
  4. 2026Evaluating Second-Order Bias of LLMs Through Epistemic Entitlementcited by 0
  5. 2026Wait, am I Being Fair? Characterizing Deductive Stereotyping and Mitigating It with Fair-GCGcited by 0
  6. 2026Position: Fairness Failure in Generative Models is an Evaluation Problemcited by 0
  7. 2026Adaptive Triggering for Bias Correction in LLM Reasoningcited by 0
  8. 2026When Who You Are Can Change the Code You Get: A Study of Persona-Induced Bias in LLM Code Generationcited by 0
  9. 2026Does Thinking Help Fairness? Reasoning Tokens Resolve Some Biases but Create Morecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026BiasTrace: Linking Reasoning Behaviours to Biased Outputs in LLMs
  2. 2026How Reasoning Shapes Social Bias in LLM-Generated Code?
  3. 2025Reasoning Towards Fairness: Mitigating Bias in Language Models through Reasoning-Guided Fine-Tuning
  4. 2024BiasAlert: A Plug-and-play Tool for Social Bias Detection in LLMs
  5. 2025Assessing Judging Bias in Large Reasoning Models: An Empirical Study
  6. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting
  7. 2025Neutralizing Bias in LLM Reasoning using Entailment Graphs
  8. 2025Towards Evaluting Fake Reasoning Bias in Language Models