Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.18970 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918042635599872 |
|---|---|
| author | Wei, Bowen Fazli, Mehrdad Zhu, Ziwei |
| author_facet | Wei, Bowen Fazli, Mehrdad Zhu, Ziwei |
| contents | Large language models (LLMs) have demonstrated impressive performance on natural language tasks, but their decision-making processes remain largely opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that humans find difficult to understand. To address these challenges, we propose \textbf{ProtoSurE}, a novel prototype-based surrogate framework that provides faithful and human-understandable explanations for LLMs. ProtoSurE trains an interpretable-by-design surrogate model that aligns with the target LLM while utilizing sentence-level prototypes as human-understandable concepts. Extensive experiments show that ProtoSurE consistently outperforms SOTA 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_18970 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning to Explain: Prototype-Based Surrogate Models for LLM Classification Wei, Bowen Fazli, Mehrdad Zhu, Ziwei Computation and Language Large language models (LLMs) have demonstrated impressive performance on natural language tasks, but their decision-making processes remain largely opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that humans find difficult to understand. To address these challenges, we propose \textbf{ProtoSurE}, a novel prototype-based surrogate framework that provides faithful and human-understandable explanations for LLMs. ProtoSurE trains an interpretable-by-design surrogate model that aligns with the target LLM while utilizing sentence-level prototypes as human-understandable concepts. Extensive experiments show that ProtoSurE consistently outperforms SOTA 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. |
| title | Learning to Explain: Prototype-Based Surrogate Models for LLM Classification |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2505.18970 |