---
id: 20260926-T0-08
title: "RLVR也能训练小型搜索智能体？可验证奖励在搜索任务上同样有效"
title_en: "RLVR Works for Small Search Agents, Study Finds"
url: https://ai.daily.yangsir.net/daily/20260926-T0-08
issue_date: 2026-09-26
publish_date: 2026-09-25T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2609.28765
---

# RLVR也能训练小型搜索智能体？可验证奖励在搜索任务上同样有效

一篇新论文探索了将可验证奖励强化学习（RLVR）应用于小型搜索智能体。RLVR此前在数学和编程等奖励信号明确的任务上表现良好，但在奖励不那么清晰的任务上是否同样有效，一直是个未解问题。该研究采用"推理-搜索"范式，将RLVR应用于开放式搜索场景，测试小模型在信息检索任务中的表现。研究结果表明，即使奖励信号不如数学题那样明确，RLVR仍能帮助小型智能体提升搜索能力。这为在资源受限场景下训练高效搜索智能体提供了新思路，开发者可以据此构建更轻量的检索增强系统。

## English Version

**RLVR Works for Small Search Agents, Study Finds**

A new paper investigates whether Reinforcement Learning with Verifiable Rewards (RLVR) can be applied to small search agents. RLVR has proven effective on tasks with clear reward signals like math and coding, but whether it works where rewards are less well-defined remains an open question. The study applies RLVR to open-ended search scenarios using a reason-over-search approach, testing small models on information retrieval tasks. Results suggest that even without crisp reward signals, RLVR can improve the search capabilities of small agents. This opens a path for training efficient search agents under resource constraints, enabling developers to build lighter retrieval-augmented systems.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2609.28765)

**详情页**：https://ai.daily.yangsir.net/daily/20260926-T0-08

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*