# Assessing Judging Bias in Large Reasoning Models: An Empirical Study

arXiv: https://arxiv.org/abs/2504.09946 · HTML: https://mlsanity.com/paper/2504.09946/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 28 of its 42 references that are in the map.
- First posted on arXiv 14 Apr 2025.

## Placed in

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

## Builds on (most cited first, 28 of 28)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [Self-Refine: Iterative Refinement with Self-Feedback](https://mlsanity.com/paper/2303.17651/) (2023)
- [Tree of Thoughts: Deliberate Problem Solving with Large Language Models](https://mlsanity.com/paper/2305.10601/) (2023)
- [A Survey on Evaluation of Large Language Models](https://mlsanity.com/paper/2307.03109/) (2023)
- [Large Language Models are not Fair Evaluators](https://mlsanity.com/paper/2305.17926/) (2023)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge](https://mlsanity.com/paper/2410.02736/) (2024)
- [JudgeBench: A Benchmark for Evaluating LLM-based Judges](https://mlsanity.com/paper/2410.12784/) (2024)
- [Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models](https://mlsanity.com/paper/2501.09686/) (2025)
- [Benchmarking Cognitive Biases in Large Language Models as Evaluators](https://mlsanity.com/paper/2309.17012/) (2023)
- [Multi-Step Reasoning with Large Language Models, a Survey](https://mlsanity.com/paper/2407.11511/) (2024)
- [Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable](https://mlsanity.com/paper/2503.00555/) (2025)
- [Long Text Generation via Adversarial Training with Leaked Information](https://mlsanity.com/paper/1709.08624/) (2017)
- [The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1](https://mlsanity.com/paper/2502.12659/) (2025)
- [LogiQA: A Challenge Dataset for Machine Reading Comprehension with Logical Reasoning](https://mlsanity.com/paper/2007.08124/) (2020)
- [Explaining Length Bias in LLM-Based Preference Evaluations](https://mlsanity.com/paper/2407.01085/) (2024)
- [RocketEval: Efficient Automated LLM Evaluation via Grading Checklist](https://mlsanity.com/paper/2503.05142/) (2025)
- [CommonGen: A Constrained Text Generation Challenge for Generative Commonsense Reasoning](https://mlsanity.com/paper/1911.03705/) (2019)
- [Distributionally Robust Receive Combining](https://mlsanity.com/paper/2401.12345/) (2024)
- [LLMs are Biased Evaluators But Not Biased for Retrieval Augmented Generation](https://mlsanity.com/paper/2410.20833/) (2024)

## Built on by (most cited first, 30 of 30)

- [One Token to Fool LLM-as-a-Judge](https://mlsanity.com/paper/2507.08794/) (2025)
- [A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models](https://mlsanity.com/paper/2509.03871/) (2025)
- [Large Language Model Reasoning Failures](https://mlsanity.com/paper/2602.06176/) (2026)
- [Whose Facts Win? LLM Source Preferences under Knowledge Conflicts](https://mlsanity.com/paper/2601.03746/) (2026)
- [Cited but Not Verified: Parsing and Evaluating Source Attribution in LLM Deep Research Agents](https://mlsanity.com/paper/2605.06635/) (2026)
- [Toward Robust LLM-Based Judges: Taxonomic Bias Evaluation and Debiasing Optimization](https://mlsanity.com/paper/2603.08091/) (2026)
- [A Mechanistic View of Authority Hierarchy in LLM Sycophancy](https://mlsanity.com/paper/2607.00415/) (2026)
- [Who Endorsed It? Measuring Authority Bias Across Expertise Levels in Language Models](https://mlsanity.com/paper/2601.13433/) (2026)
- [Beyond Gold Standards: Epistemic Ensemble of LLM Judges for Formal Mathematical Reasoning](https://mlsanity.com/paper/2506.10903/) (2025)
- [Aletheia: What Makes RLVR For Code Verifiers Tick?](https://mlsanity.com/paper/2601.12186/) (2026)
- [Fragile Thoughts: How Large Language Models Handle Chain-of-Thought Perturbations](https://mlsanity.com/paper/2603.03332/) (2026)
- [Is Your LLM-as-a-Recommender Agent Trustable? LLMs' Recommendation is Easily Hacked by Biases (Preferences)](https://mlsanity.com/paper/2603.17417/) (2026)
- [Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges](https://mlsanity.com/paper/2605.26156/) (2026)
- [Beyond Semantic Manipulation: Token-Space Attacks on Reward Models](https://mlsanity.com/paper/2604.02686/) (2026)
- [Label Effects: Shared Heuristic Reliance in Trust Assessment by Humans and LLM-as-a-Judge](https://mlsanity.com/paper/2604.05593/) (2026)
- [PulseReddit: A Novel Reddit Dataset for Benchmarking MAS in High-Frequency Cryptocurrency Trading](https://mlsanity.com/paper/2506.03861/) (2025)
- [Mitigating Translationese Bias in Multilingual LLM-as-a-Judge via Disentangled Information Bottleneck](https://mlsanity.com/paper/2603.10351/) (2026)
- [The Saturation Trap and the Subjectivity of Intervention Timing: Why Affect-Based Triggers and LLM Judges Fail to Time Interventions on Autonomous Agents](https://mlsanity.com/paper/2606.04296/) (2026)
- [AnchorBench: A Multi-Pathway Benchmark for the Anchoring Effect in LLMs](https://mlsanity.com/paper/2608.14320/) (2026)
- [Reasoning Model Is Superior LLM-Judge, Yet Suffers from Biases](https://mlsanity.com/paper/2601.03630/) (2026)
- [Themis: Training Robust Multilingual Code Reward Models for Flexible Multi-Criteria Scoring](https://mlsanity.com/paper/2605.00754/) (2026)
- [An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning Models](https://mlsanity.com/paper/2606.01462/) (2026)
- [Strategic Reflectivism In Intelligent Systems](https://mlsanity.com/paper/2505.22987/) (2025)
- [Logarithmic Scores, Power-Law Discoveries: Disentangling Measurement from Coverage in Agent-Based Evaluation](https://mlsanity.com/paper/2604.00477/) (2026)
- [Answering the Wrong Question: Reasoning Trace Inversion for Abstention in LLMs](https://mlsanity.com/paper/2604.02230/) (2026)
- [Evaluating Second-Order Bias of LLMs Through Epistemic Entitlement](https://mlsanity.com/paper/2606.17506/) (2026)
- [Chain-of-Models: Cross-Model Auditing for Bias-Robust LLM Judges](https://mlsanity.com/paper/2607.28636/) (2026)
- [The Authority Expectancy Effect in Multi-Party Conflict](https://mlsanity.com/paper/2608.08026/) (2026)
- [Beyond Correctness: Benchmarking and Aligning Response Behaviors in Hybrid-Thinking MLLMs](https://mlsanity.com/paper/2608.12781/) (2026)
- [Style Over Substance: Content-Invariant Wrappers Flip LLM Safety-Judge Verdicts](https://mlsanity.com/paper/2609.08236/) (2026)

## Abstract (arXiv)

Large Reasoning Models (LRMs) like DeepSeek-R1 and OpenAI-o1 have demonstrated remarkable reasoning capabilities, raising important questions about their biases in LLM-as-a-judge settings. We present a comprehensive benchmark comparing judging biases between LLMs and LRMs across both subjective preference-alignment datasets and objective fact-based datasets. Through investigation of bandwagon, authority, position, and distraction biases, we uncover four key findings: (1) despite their advanced reasoning capabilities, LRMs remain susceptible to the above biases; (2) LRMs demonstrate better robustness than LLMs specifically on fact-related datasets; (3) LRMs exhibit notable position bias, preferring options in later positions; and (4) we identify a novel "superficial reflection bias" where phrases mimicking reasoning (e.g., "wait, let me think...") significantly influence model judgments. To address these biases, we design and evaluate three mitigation strategies: specialized system prompts that reduce judging biases by up to 19% in preference alignment datasets and 14% in fact-related datasets, in-context learning that provides up to 27% improvement on preference tasks but shows inconsistent results on factual tasks, and a self-reflection mechanism that reduces biases by up to 10% in preference datasets and 16% in fact-related datasets, with self-reflection proving particularly effective for LRMs. Our work provides crucial insights for developing more reliable LLM-as-a-Judge frameworks, especially as LRMs become increasingly deployed as automated judges.