# Mitigating Self-Preference by Authorship Obfuscation

arXiv: https://arxiv.org/abs/2512.05379 · HTML: https://mlsanity.com/paper/2512.05379/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 18 references that are in the map.
- First posted on arXiv 5 Dec 2025.

## Builds on (most cited first, 13 of 13)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Learning to summarize from human feedback](https://mlsanity.com/paper/2009.01325/) (2020)
- [Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference](https://mlsanity.com/paper/2403.04132/) (2024)
- [Teaching Large Language Models to Self-Debug](https://mlsanity.com/paper/2304.05128/) (2023)
- [DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvature](https://mlsanity.com/paper/2301.11305/) (2023)
- [Scalable agent alignment via reward modeling: a research direction](https://mlsanity.com/paper/1811.07871/) (2018)
- [Self-critiquing models for assisting human evaluators](https://mlsanity.com/paper/2206.05802/) (2022)
- [Fast-DetectGPT: Efficient Zero-Shot Detection of Machine-Generated Text via Conditional Probability Curvature](https://mlsanity.com/paper/2310.05130/) (2023)
- [QuALITY: Question Answering with Long Input Texts, Yes!](https://mlsanity.com/paper/2112.08608/) (2021)
- [Benchmarking Foundation Models with Language-Model-as-an-Examiner](https://mlsanity.com/paper/2306.04181/) (2023)
- [Great Models Think Alike and this Undermines AI Oversight](https://mlsanity.com/paper/2502.04313/) (2025)
- [Do LLM Evaluators Prefer Themselves for a Reason?](https://mlsanity.com/paper/2504.03846/) (2025)
- [TAROT: Task-Oriented Authorship Obfuscation Using Policy Optimization Methods](https://mlsanity.com/paper/2407.21630/) (2024)

## Built on by (most cited first, 6 of 6)

- [SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning](https://mlsanity.com/paper/2607.14777/) (2026)
- [Are LLM Evaluators Really Narcissists? Sanity Checking Self-Preference Evaluations](https://mlsanity.com/paper/2601.22548/) (2026)
- [When the Reward Suite Is Leaky: A Preregistered Causal Contrast of Natural Verifier False Positives in RLVR](https://mlsanity.com/paper/2607.11022/) (2026)
- [Self- and Other-Labels Induce Bidirectional Bias in LLM Judges](https://mlsanity.com/paper/2608.18091/) (2026)
- [Sharding Prevents LLM Oversight Failures and Adversarial Exploitation](https://mlsanity.com/paper/2608.06422/) (2026)
- [Can We Trust LLM Judges: A Study of Capability-Dependent Biases and Multi-Judge Ensemble for Bias Calibration](https://mlsanity.com/paper/2609.12002/) (2026)

## Abstract (arXiv)

Language models (LMs) judges are widely used to evaluate the quality of LM outputs. Despite many advantages, LM judges display concerning biases that can impair their integrity in evaluations. One such bias is self-preference: LM judges preferring their own answers over those produced by other LMs or humans. The bias is hard to eliminate as frontier LM judges can distinguish their own outputs from those of others, even when the evaluation candidates are not labeled with their sources. In this paper, we investigate strategies to mitigate self-preference by reducing the LM judges' ability to recognize their own outputs. We apply black-box perturbations to evaluation candidates in pairwise comparison to obfuscate the authorship and reduce self-recognition. We find that perturbations as simple as synonym replacement for a few words predictably reduce self-preference. However, we also uncover fundamental challenges to eliminating the bias: when we extrapolate our perturbations to a more complete neutralization of stylistic differences between the evaluation candidates, self-preference recovers. Our findings suggest that self-recognition and self-preference can happen on many semantic levels, and complete mitigation remains challenging despite promising initial results.