Metraipolis › archive › dc_d99a2ba98442
Search-G1: Co-Evolving Representation-Based Rewards for Selective Grounded Search — arXiv:2608.07531
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.07531
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:49:44.962Z
- SHA-256 of the archived text
3a6f3e757d5c8ac618ad095a28bb83f24478c6c366ce29c7cfb0ad5ef1fd2e58- Stable URL of this page
https://metraipolis.com/documento/dc_d99a2ba98442
How to cite
Search-G1: Co-Evolving Representation-Based Rewards for Selective Grounded Search — arXiv:2608.07531 — arXiv (en), https://arxiv.org/abs/2608.07531 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d99a2ba98442 (Metraipolis archive dc_d99a2ba98442, sha256 3a6f3e757d5c8ac618ad095a28bb83f24478c6c366ce29c7cfb0ad5ef1fd2e58)
BibTeX
@misc{dcd99a2ba98442,
title = {Search-G1: Co-Evolving Representation-Based Rewards for Selective Grounded Search — arXiv:2608.07531},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.07531},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d99a2ba98442; sha256 3a6f3e757d5c8ac618ad095a28bb83f24478c6c366ce29c7cfb0ad5ef1fd2e58},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.07531. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Search-augmented language agents should retrieve external information only when necessary and ground their answers in retrieved evidence. Yet terminal correctness assigns equal credit to direct answers from parametric knowledge, evidence-grounded answers, and answers recalled after redundant search. Process supervision and LLM judges provide richer feedback but can require high-cost annotation or inference; low-cost, non-contrastive policy confidence signals do not directly identify whether retrieved evidence changes an answer. We propose Search-G1, a reward framework that couples retrieval necessity with evidence reliance through two intervention-calibrated hidden-state readouts for selective grounded search. A prompt readout predicts closed-book sufficiency, whose complement proxies policy-relative retrieval necessity; an answer-commit readout predicts evidence reliance as sensitivity to evidence deletion conditional on the realized reasoning prefix. A correctness-first reward gates evidence-reliance credit by retrieval necessity, rewards sufficient direct answers, and penalizes repeated search. After calibration, scoring needs neither process annotations nor LLM-judge inference. As policy-relative targets, trajectory distributions, and representations may drift, Search-G1 periodically regenerates targets and refits its readouts. Across four QA benchmarks and two scales, Search-G1 achieves competitive task accuracy, with stronger counterfactual evidence dependence and fewer searches in the main comparisons, complemented by independent human audits of answer support. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-11 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.07531 — 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 2026-10-19T08:49:41.823Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).