Beyond One-Size-Fits-All : Inversion Learning for Highly Effective NLG Evaluation Prompts
Transactions of the Association for Computational Linguistics · 04.05.2026 · 14 · 689-710
DOI: 10.1162/tacl.a.617 ↗Kurz gefasst
Fachartikel in der Zeitschrift Transactions of the Association for Computational Linguistics, online erschienen am 04.05.2026. Verfasst von Hanhua Hong und 5 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Evaluating natural language generation systems is challenging due to the diversity of valid outputs. While human evaluation is the gold standard, it suffers from inconsistencies, lack of standardization, and demographic biases, limiting reproducibility. LLM-based evaluators offer a scalable alternative but are highly sensitive to prompt design, where small variations can lead to significant discrepancies. In this work, we propose an inversion learning method that learns effective reverse mappings from model outputs back to their input instructions, enabling the automatic generation of highly effective, model-specific evaluation prompts. Our method requires only a single evaluation sample and eliminates the need for time-consuming manual prompt engineering, thereby improving both efficiency and robustness. Our work contributes toward a new direction for more robust and efficient LLM-based evaluation.
Zitieren
APA
Hong, H., Xiao, C., Wang, Y., Liu, Y., Rong, W., & Lin, C. (2026). Beyond One-Size-Fits-All : Inversion Learning for Highly Effective NLG Evaluation Prompts. Transactions of the Association for Computational Linguistics, 14, 689-710. https://doi.org/10.1162/tacl.a.617
Vancouver
Hong H, Xiao C, Wang Y, Liu Y, Rong W, Lin C. Beyond One-Size-Fits-All : Inversion Learning for Highly Effective NLG Evaluation Prompts. Transactions of the Association for Computational Linguistics. 2026;14:689-710. doi:10.1162/tacl.a.617
BibTeX
@article{Hong2026,
author = {Hong, Hanhua and Xiao, Chenghao and Wang, Yang and Liu, Yiqi and Rong, Wenge and Lin, Chenghua},
title = {Beyond One-Size-Fits-All : Inversion Learning for Highly Effective NLG Evaluation Prompts},
journal = {Transactions of the Association for Computational Linguistics},
year = {2026},
volume = {14},
pages = {689--710},
doi = {10.1162/tacl.a.617}
}
EUVIMED