RAISE: Enhancing Scientific Reasoning in LLMs via Step-by-Step Retrieval
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_ | 1866908898578923520 |
|---|---|
| author | Oh, Minhae Kim, Jeonghye Lee, Nakyung Seo, Donggeon Kim, Taeuk Lee, Jungwoo |
| author_facet | Oh, Minhae Kim, Jeonghye Lee, Nakyung Seo, Donggeon Kim, Taeuk Lee, Jungwoo |
| contents | Scientific reasoning requires not only long-chain reasoning processes, but also knowledge of domain-specific terminologies and adaptation to updated findings. To deal with these challenges for scientific reasoning, we introduce RAISE, a step-by-step retrieval-augmented framework which retrieves logically relevant documents from in-the-wild corpus. RAISE is divided into three steps: problem decomposition, logical query generation, and logical retrieval. We observe that RAISE consistently outperforms other baselines on scientific reasoning benchmarks. We analyze that unlike other baselines, RAISE retrieves documents that are not only similar in terms of the domain knowledge, but also documents logically more relevant. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_08625 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RAISE: Enhancing Scientific Reasoning in LLMs via Step-by-Step Retrieval Oh, Minhae Kim, Jeonghye Lee, Nakyung Seo, Donggeon Kim, Taeuk Lee, Jungwoo Computation and Language Scientific reasoning requires not only long-chain reasoning processes, but also knowledge of domain-specific terminologies and adaptation to updated findings. To deal with these challenges for scientific reasoning, we introduce RAISE, a step-by-step retrieval-augmented framework which retrieves logically relevant documents from in-the-wild corpus. RAISE is divided into three steps: problem decomposition, logical query generation, and logical retrieval. We observe that RAISE consistently outperforms other baselines on scientific reasoning benchmarks. We analyze that unlike other baselines, RAISE retrieves documents that are not only similar in terms of the domain knowledge, but also documents logically more relevant. |
| title | RAISE: Enhancing Scientific Reasoning in LLMs via Step-by-Step Retrieval |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2506.08625 |