Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Jianghao, Tang, Feilong, Li, Yulong, Hu, Ming, Xue, Haochen, Jameel, Shoaib, Xie, Yutong, Razzak, Imran
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2505.18283
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909621927542784
author Wu, Jianghao
Tang, Feilong
Li, Yulong
Hu, Ming
Xue, Haochen
Jameel, Shoaib
Xie, Yutong
Razzak, Imran
author_facet Wu, Jianghao
Tang, Feilong
Li, Yulong
Hu, Ming
Xue, Haochen
Jameel, Shoaib
Xie, Yutong
Razzak, Imran
contents Recent advances such as Chain-of-Thought prompting have significantly improved large language models (LLMs) in zero-shot medical reasoning. However, prompting-based methods often remain shallow and unstable, while fine-tuned medical LLMs suffer from poor generalization under distribution shifts and limited adaptability to unseen clinical scenarios. To address these limitations, we present TAGS, a test-time framework that combines a broadly capable generalist with a domain-specific specialist to offer complementary perspectives without any model fine-tuning or parameter updates. To support this generalist-specialist reasoning process, we introduce two auxiliary modules: a hierarchical retrieval mechanism that provides multi-scale exemplars by selecting examples based on both semantic and rationale-level similarity, and a reliability scorer that evaluates reasoning consistency to guide final answer aggregation. TAGS achieves strong performance across nine MedQA benchmarks, boosting GPT-4o accuracy by 13.8%, DeepSeek-R1 by 16.8%, and improving a vanilla 7B model from 14.1% to 23.9%. These results surpass several fine-tuned medical LLMs, without any parameter updates. The code will be available at https://github.com/JianghaoWu/TAGS.
format Preprint
id arxiv_https___arxiv_org_abs_2505_18283
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TAGS: A Test-Time Generalist-Specialist Framework with Retrieval-Augmented Reasoning and Verification
Wu, Jianghao
Tang, Feilong
Li, Yulong
Hu, Ming
Xue, Haochen
Jameel, Shoaib
Xie, Yutong
Razzak, Imran
Computation and Language
Artificial Intelligence
Multiagent Systems
I.2.7
Recent advances such as Chain-of-Thought prompting have significantly improved large language models (LLMs) in zero-shot medical reasoning. However, prompting-based methods often remain shallow and unstable, while fine-tuned medical LLMs suffer from poor generalization under distribution shifts and limited adaptability to unseen clinical scenarios. To address these limitations, we present TAGS, a test-time framework that combines a broadly capable generalist with a domain-specific specialist to offer complementary perspectives without any model fine-tuning or parameter updates. To support this generalist-specialist reasoning process, we introduce two auxiliary modules: a hierarchical retrieval mechanism that provides multi-scale exemplars by selecting examples based on both semantic and rationale-level similarity, and a reliability scorer that evaluates reasoning consistency to guide final answer aggregation. TAGS achieves strong performance across nine MedQA benchmarks, boosting GPT-4o accuracy by 13.8%, DeepSeek-R1 by 16.8%, and improving a vanilla 7B model from 14.1% to 23.9%. These results surpass several fine-tuned medical LLMs, without any parameter updates. The code will be available at https://github.com/JianghaoWu/TAGS.
title TAGS: A Test-Time Generalist-Specialist Framework with Retrieval-Augmented Reasoning and Verification
topic Computation and Language
Artificial Intelligence
Multiagent Systems
I.2.7
url https://arxiv.org/abs/2505.18283