[论文] Pterodactyl: Pushing the Limits of Efficient Automated Translation Eva...

论文概要

研究领域: NLP 作者: Evelyn Y. Garland, Carola F. Berger 发布时间: 2026-11-10 arXiv: 2611.14480

中文摘要

我们介绍 1020 Group 在 WMT26 自动质量评估共享任务上的提交系统 Pterodactyl。该系统专为子任务 2 和 3 设计，即片段级质量分数预测与无错片段检测。Pterodactyl 是一个高度可定制、超低成本的系统，针对翻译分诊与数据筛选中的二分类优化。它基于 LLM-as-a-judge 方法，提示工程与模型选择由 ROC 分析引导。我们还开发了若干专门适配子任务 3 的 Pterodactyl 变体。

原文摘要

In this paper, we present 1020 Group's submission to the WMT26 shared task on automated quality evaluation. Our system, Pterodactyl, was engineered specifically for subtasks 2 and 3, that is, segment-level quality score prediction and the detection of error-free segments. Pterodactyl is a highly customizable, ultra-low cost system optimized for binary classification in translation triage and data curation. It is based on an LLM-as-a-judge approach, with prompt engineering and model selection guided by ROC analysis. We also developed several variants of Pterodactyl that were specifically adapted for subtasks 3.

#论文 #arXiv #NLP #小凯