The paper introduces the "Prompt Lottery" phenomenon, where paraphrastic rephrasings of judge prompts in LLM-as-a-Judge (LLMaJ) lead to inconsistent judgments and can even reverse model rankings. A large-scale study across various tasks, judge families, and model backbones demonstrates that these prompt variations significantly alter evaluation outcomes. A user study shows that LLMaJ judges are 6–9 times more sensitive to prompt rephrasings than humans, and that aggregating multiple prompts improves robustness by focusing on consistent judgments.
By Ilija Subasic, Brett Levert, Andrew Rabinovich, Sajjad Pakdamansavoji, Zhao Chen
arXiv:2608. 15535v1 Announce Type: cross Abstract: We present L3Cube-IndicQuest v2, a large-scale gold-standard multilingual question-answering benchmark for evaluating the India-specific factual knowledge of Large Language Models (LLMs).
By Rinit Jain, Tirthraj Mahajan, Advait Joshi, Raviraj Joshi
arXiv:2606. 10327v1 Announce Type: cross Abstract: Automated Essay Scoring (AES) systems must judge interdependent discourse elements (e.
By Ali Keramati, Mark Warschauer
The paper demonstrates that large language model (LLM) evaluators, whether reward‑model based or prompted LLM‑as‑a‑Judge, exhibit significant language bias in multilingual settings. Experiments with semantically identical instruction‑response pairs across 23 languages reveal that lower‑resource languages receive higher scores, a bias that persists across eight open‑weight evaluators and is not detectable by standard pairwise accuracy metrics. The authors link the bias to model uncertainty and language identity, showing it cannot be explained by content difficulty alone.
By Ej Zhou, Lucas Resck, Zheng Hui, Anna Korhonen
The paper investigates whether language models’ self-reported confidence is meaningful without additional training. By evaluating three training‑free signals—direct verbalization, post‑hoc probability estimates, and agreement across multiple generations—on 100 TriviaQA questions, the authors find that direct verbalization alone achieves high AUROC scores (0.956 and 0.937) for correctness prediction, while agreement-based methods perform noticeably worse. Re‑eliciting confidence for the same answers shows modest score shifts and occasional decision flips, and an audit of biography claims reveals only a small confidence gap between supported and contradicted statements.
By Lukas Meyer, Sofia Rossi, Wei Chen, Thomas Laurent, Yiming Li
The study investigates how different prompting strategies affect large language model (LLM) judges across multiple languages and tasks. Using a benchmark of 9 prompting strategies, 4 open-weight judges, 27 languages, and 4 tasks, the authors collected over 1 million pointwise scores validated against human judgments. Results show that prompting effectiveness is highly dependent on the judge model, can reverse direction across judges, and becomes less stable in low‑resource languages, indicating that prompt choice must be tailored to the specific evaluation setting.
By Asal Mehrabi, Mohammad Aliannejadi, Clemencia Siro
arXiv:2607. 19243v1 Announce Type: cross Abstract: Although Large Language Models (LLMs) demonstrate remarkable multilingual fluency, their internal knowledge representations remain disproportionately biased toward high-resource languages.
By Alexander Manev
The study compares four approaches for Chinese sentence-level metaphor identification: BERT fine‑tuning, QLoRA-based large language model fine‑tuning, zero‑shot LLM prompting, and zero‑shot prompting with an expert‑informed procedural Skill. Results show that fine‑tuning yields the highest accuracy on the native test set, while the Skill‑based zero‑shot method provides the most stable performance across three datasets, achieving the highest external floor and the smallest performance range. Adding the Skill reduces false positives on one dataset but increases false negatives on others, indicating a trade‑off between precision and recall.
By Yufeng Wu, Meichun Liu
arXiv:2609.13824v1 Announce Type: new
Abstract: Large language models (LLMs) are increasingly used to evaluate the responses of other language models. This approach, known as LLM-as-a-Judge, is faste...
By Aakash Kumar Tiwari
arXiv:2609.13773v1 Announce Type: new
Abstract: LLM-as-a-Judge evaluators are increasingly used to score open-ended generation, yet a judge's correlation with human ratings on its development set may...
By Ruichen Zheng, Yihe Wang, Fabrice Y Harel-Canada, Sara Khosravi, Zeynep Senahan Yildiz, Amit Sahai, Nanyun Peng
arXiv:2607. 13189v1 Announce Type: cross Abstract: We present RAGthoven, our system for SemEval-2026 Task 1 (MWAHAHA), Subtask A (multilingual constrained humor generation in English, Spanish, and Chinese).
By Marek \v{S}uppa, Vikt\'oria Ondrejov\'a, Lucia Ganajov\'a, Gregor Karetka, Daniel Skala
arXiv:2608.28626v1 Announce Type: cross
Abstract: Large language models (LLMs) are increasingly used to generate peer reviews, prompting examination of their capacity for critical evaluation. This st...
By Emad Alharbi