Metraipolis › archive › dc_f5249dc93a06
Route-and-Generate: KL-Regularized Router Improvement for Reasoning Routing in LLMs — arXiv:2610.30351
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.30351
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-25T08:32:24.129Z
- SHA-256 of the archived text
9159d8d782bd32f01014e3c2d26f648d68e5ff2de15b7d38ca8e708ae4618ae3- Stable URL of this page
https://metraipolis.com/documento/dc_f5249dc93a06
How to cite
Route-and-Generate: KL-Regularized Router Improvement for Reasoning Routing in LLMs — arXiv:2610.30351 — arXiv (en), https://arxiv.org/abs/2610.30351 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f5249dc93a06 (Metraipolis archive dc_f5249dc93a06, sha256 9159d8d782bd32f01014e3c2d26f648d68e5ff2de15b7d38ca8e708ae4618ae3)
BibTeX
@misc{dcf5249dc93a06,
title = {Route-and-Generate: KL-Regularized Router Improvement for Reasoning Routing in LLMs — arXiv:2610.30351},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.30351},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f5249dc93a06; sha256 9159d8d782bd32f01014e3c2d26f648d68e5ff2de15b7d38ca8e708ae4618ae3},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30351. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models often apply complex reasoning indiscriminately, leading to overthinking on simple tasks and underthinking on harder ones. We present Route-and-Generate (Rouge), a hierarchical training framework that enables LLMs to adaptively decide when and how much reasoning is required for each input. Our approach decouples reasoning effort allocation from answer generation by introducing a routing policy that allocates a fixed sampling budget across multiple reasoning formats, while the base model is optimized with GRPO using per-sample rewards. To promote efficient reasoning, we employ a rank-based reward that favors shorter correct solutions, and update the routing policy via a closed-form KL-regularized policy improvement step distilled into a neural router. This design encourages concise answers on easy problems while reserving longer reasoning for harder ones, achieving improved accuracy-efficiency trade-offs across multiple reasoning benchmarks. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.30351 — 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 2026-10-25T08:31:12.982Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).