Rule-Based Explanations for Retrieval-Augmented LLM Systems
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918396711403520 |
|---|---|
| author | Rorseth, Joel Godfrey, Parke Golab, Lukasz Srivastava, Divesh Szlichta, Jarek |
| author_facet | Rorseth, Joel Godfrey, Parke Golab, Lukasz Srivastava, Divesh Szlichta, Jarek |
| contents | If-then rules are widely used to explain machine learning models; e.g., "if employed = no, then loan application = rejected." We present the first proposal to apply rules to explain the emerging class of large language models (LLMs) with retrieval-augmented generation (RAG). Since RAG enables LLM systems to incorporate retrieved information sources at inference time, rules linking the presence or absence of sources can explain output provenance; e.g., "if a Times Higher Education ranking article is retrieved, then the LLM ranks Oxford first." To generate such rules, a brute force approach would probe the LLM with all source combinations and check if the presence or absence of any sources leads to the same output. We propose optimizations to speed up rule generation, inspired by Apriori-like pruning from frequent itemset mining but redefined within the scope of our novel problem. We conclude with qualitative and quantitative experiments demonstrating our solutions' value and efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_22689 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Rule-Based Explanations for Retrieval-Augmented LLM Systems Rorseth, Joel Godfrey, Parke Golab, Lukasz Srivastava, Divesh Szlichta, Jarek Computation and Language If-then rules are widely used to explain machine learning models; e.g., "if employed = no, then loan application = rejected." We present the first proposal to apply rules to explain the emerging class of large language models (LLMs) with retrieval-augmented generation (RAG). Since RAG enables LLM systems to incorporate retrieved information sources at inference time, rules linking the presence or absence of sources can explain output provenance; e.g., "if a Times Higher Education ranking article is retrieved, then the LLM ranks Oxford first." To generate such rules, a brute force approach would probe the LLM with all source combinations and check if the presence or absence of any sources leads to the same output. We propose optimizations to speed up rule generation, inspired by Apriori-like pruning from frequent itemset mining but redefined within the scope of our novel problem. We conclude with qualitative and quantitative experiments demonstrating our solutions' value and efficiency. |
| title | Rule-Based Explanations for Retrieval-Augmented LLM Systems |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.22689 |