BatchGEMBA: Token-Efficient Machine Translation Evaluation with Batched Prompting and Prompt Compression

Abstract

Recent advancements in Large Language Model (LLM)-based Natural Language Generation evaluation have largely focused on single-example prompting, resulting in significant token overhead and computational inefficiencies. In this work, we introduce BatchGEMBA-MQM, a framework that integrates batched prompting with the GEMBA-MQM metric for machine translation evaluation. Our approach aggregates multiple translation examples into a single prompt, reducing token usage by 2–4 times (depending on the batch size) relative to single-example prompting. Evaluations across several LLMs (GPT-4o, GPT-4o-mini, Mistral Small, Phi4, and CommandR7B) and varying batch sizes reveal that while batching generally negatively affects quality (but sometimes not substantially). Furthermore, we propose a batching-aware prompt compression model that achieves an additional token reduction of 13–15% on average while also showing ability to help mitigate batching-induced quality degradation. For instance, GPT-4o retains over 90% of its baseline performance at a batch size of 4 when compression is applied, compared to a 44.6% drop without compression. We plan to release our code and trained models at https://github.com/NL2G/batchgemba to support future research in this domain.

Ähnliche Arbeiten