Metraipolis › archive › dc_f688d059cde7

ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment — arXiv:2608.05102

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.05102
License (as archived)
CC0-1.0
Archived
2026-10-31T11:13:18.254Z
SHA-256 of the archived text
0c53c80fc341ccb6e462f5f3a36bbc89d2c470a560ddc6ed55dc570853173db5
Stable URL of this page
https://metraipolis.com/documento/dc_f688d059cde7

How to cite

ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment — arXiv:2608.05102 — arXiv (en), https://arxiv.org/abs/2608.05102 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f688d059cde7 (Metraipolis archive dc_f688d059cde7, sha256 0c53c80fc341ccb6e462f5f3a36bbc89d2c470a560ddc6ed55dc570853173db5)
BibTeX
@misc{dcf688d059cde7,
  title = {ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment — arXiv:2608.05102},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.05102},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f688d059cde7; sha256 0c53c80fc341ccb6e462f5f3a36bbc89d2c470a560ddc6ed55dc570853173db5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05102. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-horizon search agents must make multiple sequential actions (steps) to search, retrieve, verify, and integrate evidence to reach a final answer. However, existing methods for training these agents typically treat all steps within a trajectory uniformly during both supervised fine-tuning (SFT) and reinforcement learning (RL), failing to distinguish useful actions from erroneous or redundant ones. In this paper, we propose Answer-Backtracked Credit Assignment (ABC), a fine-grained credit assignment framework for training long-horizon search agents by converting sparse trajectory-level outcomes into dense step-level supervision that rewards useful actions (even in failed trajectories) while suppressing erroneous or redundant actions. Specifically, given a potentially obscure query and its corresponding ground-truth answer, ABC first performs Answer-Backtracked Clue Recovery, which traces back from the answer to recover intermediate clues required to solve the question. It then applies Clue-Anchored Step Scoring to evaluate each search step against these clues, converting sparse binary outcome supervision into dense step-level rewards. Based on these rewards, we develop ABC-SFT, which reweights the loss of each turn, and ABC-GRPO, which uses the step-level scores as rewards in GRPO. Building on this framework, we train ABSeeker based on Qwen3.5-4B with only 8.5k examples. ABSeeker achieves 37.3% on BrowseComp and 39.1% on BrowseComp-ZH. With context management, the scores further improve to 55.3% and 52.9%, respectively, significantly outperforming same-scale (4B) agents and even matching the performance of larger ones (approximately 30B). These results demonstrate the effectiveness of answer-backtracked step-level credit assignment for training long-horizon search agents.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-06 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.05102 — 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 2026-10-31T11:02:13.441Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).