# Bias in the Loop: Auditing LLM-as-a-Judge for Software Engineering

arXiv: https://arxiv.org/abs/2604.16790 · HTML: https://mlsanity.com/paper/2604.16790/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 35 references that are in the map.
- First posted on arXiv 18 Apr 2026.

## Builds on (most cited first, 25 of 25)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [A Survey on Large Language Models for Code Generation](https://mlsanity.com/paper/2406.00515/) (2024)
- [From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline](https://mlsanity.com/paper/2406.11939/) (2024)
- [LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods](https://mlsanity.com/paper/2412.05579/) (2024)
- [Training Software Engineering Agents and Verifiers with SWE-Gym](https://mlsanity.com/paper/2412.21139/) (2024)
- [WebThinker: Empowering Large Reasoning Models with Deep Research Capability](https://mlsanity.com/paper/2504.21776/) (2025)
- [JudgeBench: A Benchmark for Evaluating LLM-based Judges](https://mlsanity.com/paper/2410.12784/) (2024)
- [Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges](https://mlsanity.com/paper/2406.12624/) (2024)
- [Large Language Model-Based Agents for Software Engineering: A Survey](https://mlsanity.com/paper/2409.02977/) (2024)
- [SWE-Search: Enhancing Software Agents with Monte Carlo Tree Search and Iterative Refinement](https://mlsanity.com/paper/2410.20285/) (2024)
- [Can LLMs Replace Human Evaluators? An Empirical Study of LLM-as-a-Judge in Software Engineering](https://mlsanity.com/paper/2502.06193/) (2025)
- [A Survey on Large Language Models for Software Engineering](https://mlsanity.com/paper/2312.15223/) (2023)
- [ICE-Score: Instructing Large Language Models to Evaluate Code](https://mlsanity.com/paper/2304.14317/) (2023)
- [Agentic Software Engineering: Foundational Pillars and a Research Roadmap](https://mlsanity.com/paper/2509.06216/) (2025)
- [CodeJudgeBench: Benchmarking LLM-as-a-Judge for Coding Tasks](https://mlsanity.com/paper/2507.10535/) (2025)
- [CodeJudge-Eval: Can Large Language Models be Good Judges in Code Understanding?](https://mlsanity.com/paper/2408.10718/) (2024)
- [Bias Unveiled: Investigating Social Bias in LLM-Generated Code](https://mlsanity.com/paper/2411.10351/) (2024)
- [Agentic Rubrics as Contextual Verifiers for SWE Agents](https://mlsanity.com/paper/2601.04171/) (2026)
- [PatchPilot: A Cost-Efficient Software Engineering Agent with Early Attempts on Formal Verification](https://mlsanity.com/paper/2502.02747/) (2025)
- [Don't Judge Code by Its Cover: Exploring Biases in LLM Judges for Code Evaluation](https://mlsanity.com/paper/2505.16222/) (2025)
- [On the Effectiveness of LLM-as-a-judge for Code Generation and Summarization](https://mlsanity.com/paper/2507.16587/) (2025)
- [From Code to Courtroom: LLMs as the New Software Judges](https://mlsanity.com/paper/2503.02246/) (2025)
- [Social Bias Benchmark for Generation: A Comparison of Generation and QA-Based Evaluations](https://mlsanity.com/paper/2503.06987/) (2025)
- [Efficiency vs. Alignment: Investigating Safety and Fairness Risks in Parameter-Efficient Fine-Tuning of LLMs](https://mlsanity.com/paper/2511.00382/) (2025)

## Built on by (most cited first, 4 of 4)

- [Scaffold, Not Vocabulary? A Controlled, Two-Tier, Pre-Registered Study of a Popperian Code-Generation Skill](https://mlsanity.com/paper/2606.06454/) (2026)
- [PaperJury: Due-Process Review for Bounded LaTeX Revision](https://mlsanity.com/paper/2606.16322/) (2026)
- [Groundability, Not Scale Alone: When Weak Reviewers Can Audit Strong Coding Agents](https://mlsanity.com/paper/2610.01023/) (2026)
- [Where Did the Repair First Go Wrong? Localizing the Origins of Silent Failures in Agentic Vulnerability Repair](https://mlsanity.com/paper/2610.06163/) (2026)

## Abstract (arXiv)

Large Language Models are increasingly used as judges to evaluate code artifacts when exhaustive human review or executable test coverage is unavailable. LLM-judge is increasingly relevant in agentic software engineering workflows, where it can help rank candidate solutions and guide patch selection. While attractive for scale, current practice lacks a principled account of reliability and bias: repeated evaluations of the same case can disagree; small prompt edits can swing outcomes; and seemingly semantics-preserving, human-equivalent perturbations may elicit divergent verdicts. This paper studies LLM-as-a-Judge for code through a measurement-first lens. We analyze two pointwise judging regimes across code generation, code repair task, and test generation, and we systematically probe prompt-induced biases. Our study considers difficulty levels for repeated runs and controlled prompt interventions that isolate one presentation cue at a time, and it evaluates judges using consistency and sensitivity to bias. We find that judge decisions are highly sensitive to prompt biases even when the underlying code snippet is unchanged. Across all three tasks, several biases systematically shift preferences toward the option favored by the prompt, improving accuracy when that option aligns with the gold answer but substantially reducing it otherwise. In some settings, these effects are large enough to change task-level conclusions and alter relative model rankings. These findings show that reported judge performance may reflect prompt artifacts rather than stable assessment ability, posing a direct threat to the validity and reproducibility of code evaluation. We therefore argue that LLM-as-a-Judge studies should report bias sensitivity alongside accuracy and incorporate explicit controls to support more trustworthy model comparison in software engineering.