EUVIMEDEUVIMED entdecken
Crossref

Making Sense of LLM Decisions: A Prototype-based Framework for Explainable Classification

Bowen Wei, Mehrdad Fazli, Ziwei Zhu

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (32) · 26814-26822

DOI: 10.1609/aaai.v40i32.39892 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Bowen Wei und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language models have demonstrated impressive performance on natural language tasks, but their decision-making processes remain opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that are difficult for humans to understand. To address these challenges, we propose ProtoSurE, a novel prototype-based surrogate framework that provides faithful and understandable explanations for LLMs. ProtoSurE trains an interpretable-by-design surrogate model that aligns with the target LLM while utilizing sentence-level prototypes as understandable concepts. Extensive experiments show that ProtoSurE consistently outperforms state-of-the-art explanation methods across diverse LLMs and datasets. Importantly, ProtoSurE demonstrates strong data efficiency, requiring relatively few training examples to achieve good performance, making it practical for real-world applications.

Zitieren

APA

Wei, B., Fazli, M., & Zhu, Z. (2026). Making Sense of LLM Decisions: A Prototype-based Framework for Explainable Classification. Proceedings of the AAAI Conference on Artificial Intelligence, 40(32), 26814-26822. https://doi.org/10.1609/aaai.v40i32.39892

Vancouver

Wei B, Fazli M, Zhu Z. Making Sense of LLM Decisions: A Prototype-based Framework for Explainable Classification. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(32):26814-26822. doi:10.1609/aaai.v40i32.39892

BibTeX
@article{Wei2026,
  author = {Wei, Bowen and Fazli, Mehrdad and Zhu, Ziwei},
  title = {Making Sense of LLM Decisions: A Prototype-based Framework for Explainable Classification},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {32},
  pages = {26814--26822},
  doi = {10.1609/aaai.v40i32.39892}
}
Kostenlos starten →