Metraipolis › archive › dc_421f260fc4ad
Disagree to Explore, Agree to Commit: Routing-Guided Test-Time Scaling for Software Agents — arXiv:2608.22191
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.22191
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T07:41:20.814Z
- SHA-256 of the archived text
5e5977b2d4057b93394bdd947a5af7230371ebf87cbbc73f48680c0c60181e58- Stable URL of this page
https://metraipolis.com/documento/dc_421f260fc4ad
How to cite
Disagree to Explore, Agree to Commit: Routing-Guided Test-Time Scaling for Software Agents — arXiv:2608.22191 — arXiv (en), https://arxiv.org/abs/2608.22191 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_421f260fc4ad (Metraipolis archive dc_421f260fc4ad, sha256 5e5977b2d4057b93394bdd947a5af7230371ebf87cbbc73f48680c0c60181e58)
BibTeX
@misc{dc421f260fc4ad,
title = {Disagree to Explore, Agree to Commit: Routing-Guided Test-Time Scaling for Software Agents — arXiv:2608.22191},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.22191},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_421f260fc4ad; sha256 5e5977b2d4057b93394bdd947a5af7230371ebf87cbbc73f48680c0c60181e58},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.22191. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Software-engineering agents solve repository-level tasks through long, stochastic tool-use trajectories, and repeated attempts often find fixes missed by one run. Test-time scaling is difficult because patches lack canonical answer forms, while sibling actions from a shared prefix are correlated. We study whether native MoE router traces can guide steering and selection without an external judge or selection-time test execution. Our analysis shows that routing provides a robust behavioral role signal; token-granular readouts and decision-matched comparison sets turn it into effective control. We therefore introduce Risa (Routing-Informed Steering and Arbitration): within trajectories, routing encourages diverse exploration and controlled convergence during patch commitment; across separately sampled trajectories, agreement at informative patch positions selects a final candidate. We evaluate on SWE-bench Verified using open-weight sparse MoE agents across scales and reasoning-effort settings. Risa's routing arbitration raises the macro-average resolved rate from 44.9% under uniform sampling to 48.2% on the gpt-oss family, matching text consensus without answer-string matching, and it transfers to Qwen3.6, where it improves on uniform choice and matches text consensus on the full 500-task benchmark. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2608.22191 — 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 2026-10-18T07:37:52.710Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).