Metraipolis › archive › dc_10ec9539bbbd

PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents — arXiv:2609.35816

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.35816
License (as archived): CC0-1.0
Archived: 2026-10-07T16:24:46.001Z
SHA-256 of the archived text: e1fa2092a65a5502bd1349b6c2c820c600eb18215670f1287988adc697d7c2cf
Stable URL of this page: https://metraipolis.com/documento/dc_10ec9539bbbd

How to cite

PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents — arXiv:2609.35816 — arXiv (en), https://arxiv.org/abs/2609.35816 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_10ec9539bbbd (Metraipolis archive dc_10ec9539bbbd, sha256 e1fa2092a65a5502bd1349b6c2c820c600eb18215670f1287988adc697d7c2cf)

BibTeX

@misc{dc10ec9539bbbd,
  title = {PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents — arXiv:2609.35816},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35816},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_10ec9539bbbd; sha256 e1fa2092a65a5502bd1349b6c2c820c600eb18215670f1287988adc697d7c2cf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35816. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model search agents are often trained with synthetic questions whose difficulty is increased through larger evidence graphs, additional hops, and longer trajectories. These global properties, however, are only indirect proxies for the local retrieval capabilities required during search. To address this mismatch, we introduce latent anchor reasoning, which consists of resolving an unnamed retrieval anchor from descriptive specifications and transferring the recovered anchor into a subsequent information demand. This primitive retrieval unit decomposes deep search into chains of coupled operations and organizes question construction around anchor resolution and relation transfer, without prescribing a canonical search path. Based on this formulation, we propose PrimeSeeker, a capability-oriented framework that constructs web-grounded anchor structures and jointly derives a question and a reference evidence skeleton. The skeleton preserves supporting evidence from construction and guides expert generation through extractive highlights of current tool observations. These highlights are removed before supervised fine-tuning, while the skeleton is subsequently reused to audit reference-step coverage for reinforcement-learning rewards. We construct 9,221 expert trajectories, training a 30B search agent. Across five deep-search benchmarks, PrimeSeeker achieves strong performance, while reference-step optimization further improves the supervised policy. The resulting trajectories exhibit low retrieval redundancy, and fixed-budget evaluation shows strong solution coverage with substantially fewer tool calls than long-horizon systems.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.35816 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T15:54:53.716Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).