Retrieval, Reward, and Training Protocols: What Matters in Training Search Agents?
Search agents powered by large language models can autonomously decompose queries, retrieve information, and synthesize answers through multi-step reasoning. However, the rapid growth of training methods has outpaced controlled comparison: existing works differ in retrieval corpora, reward designs, and training protocols, making it unclear what actually drives improvements. We present a controlled empirical study that isolates three under-explored dimensions of search agent training. First, we identify a critical data-coverage issue in the widely used Wikipedia 2018 corpus and show that correcting it alone yields larger gains than the differences between training algorithms. Second, we sy...
| Authors | Yibo Zhao, Zichen Ding, Jiayi Wu et al. |
|---|---|
| Primary category | cs.CL |
| Published | 27 May 2026 |
| Last updated at source | 29 May 2026 |
Connected in the knowledge graph
- ybyby624/wiki-fixedDataset · cited by
Models and datasets referencing this paper
- ybyby624/wiki-fixedDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.