# Multilingual Prompt Localization for Agent-as-a-Judge: Language and Backbone Sensitivity in Requirement-Level Evaluation

arXiv: https://arxiv.org/abs/2604.04532 · HTML: https://mlsanity.com/paper/2604.04532/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 22 of its 29 references that are in the map.
- First posted on arXiv 6 Apr 2026.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 22 of 22)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework](https://mlsanity.com/paper/2308.00352/) (2023)
- [Annotation Artifacts in Natural Language Inference Data](https://mlsanity.com/paper/1803.02324/) (2018)
- [MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering](https://mlsanity.com/paper/2410.07095/) (2024)
- [Agent-as-a-Judge: Evaluate Agents with Agents](https://mlsanity.com/paper/2410.10934/) (2024)
- [Data Contamination: From Memorization to Exploitation](https://mlsanity.com/paper/2203.08242/) (2022)
- [FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets](https://mlsanity.com/paper/2307.10928/) (2023)
- [MASSIVE: A 1M-Example Multilingual Natural Language Understanding Dataset with 51 Typologically-Diverse Languages](https://mlsanity.com/paper/2204.08582/) (2022)
- [IndicGenBench: A Multilingual Benchmark to Evaluate Generation Capabilities of LLMs on Indic Languages](https://mlsanity.com/paper/2404.16816/) (2024)
- [McEval: Massively Multilingual Code Evaluation](https://mlsanity.com/paper/2406.07436/) (2024)
- [MM-Eval: A Multilingual Meta-Evaluation Benchmark for LLM-as-a-Judge and Reward Models](https://mlsanity.com/paper/2410.17578/) (2024)
- [mHumanEval -- A Multilingual Benchmark to Evaluate Large Language Models for Code Generation](https://mlsanity.com/paper/2410.15037/) (2024)
- [How Reliable is Multilingual LLM-as-a-Judge?](https://mlsanity.com/paper/2505.12201/) (2025)
- [GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks Through Code Repository Leveraging](https://mlsanity.com/paper/2508.18993/) (2025)
- [Cross-Lingual Auto Evaluation for Assessing Multilingual LLMs](https://mlsanity.com/paper/2410.13394/) (2024)
- [SWE-Compass: Towards Unified Evaluation of Agentic Coding Abilities for Large Language Models](https://mlsanity.com/paper/2511.05459/) (2025)
- [Time To Impeach LLM-as-a-Judge: Programs are the Future of Evaluation](https://mlsanity.com/paper/2506.10403/) (2025)
- [OmniCode: A Benchmark for Evaluating Software Engineering Agents](https://mlsanity.com/paper/2602.02262/) (2026)
- [Cross-Lingual LLM-Judge Transfer via Evaluation Decomposition](https://mlsanity.com/paper/2603.18557/) (2026)

## Built on by (most cited first, 2 of 2)

- [Rank Reversal in Multilingual LLM Judges: A Label-Free Double-Centering Calibrator](https://mlsanity.com/paper/2608.22432/) (2026)
- [When Agent Evaluation Becomes Unreliable: An Empirical Study of Oracle Failures and Configuration-Dependent Measurement in an LLM Agent Benchmark](https://mlsanity.com/paper/2610.16067/) (2026)

## Abstract (arXiv)

Evaluation language is typically treated as a fixed English default in agentic code benchmarks, yet we show that changing the judge's language can invert backbone rankings. We localize the Agent-as-a-Judge prompt stack to five typologically diverse languages (English, Arabic, Turkish, Chinese, Hindi) and evaluate 55 DevAI development tasks across three developer-agent frameworks and six judge backbones, totaling 4950 judge runs. The central finding is that backbone and language interact: GPT-4o achieves the highest satisfaction in English (44.72%), while Gemini leads in Arabic (51.72%, p<0.001 vs. GPT-4o) and Hindi (53.22%). No single backbone dominates across all languages, and inter-backbone agreement on individual requirement judgments is modest (Fleiss' kappa ≤ 0.231). A controlled ablation further shows that localizing judge-side instructions, not just benchmark content, can be decisive: Hindi satisfaction drops from 42.8% to 23.2% under partial localization. These results indicate that language should be treated as an explicit evaluation variable in agentic benchmarks. Full requirement-level judgments and runtime statistics are released for reproducibility.