Metraipolis › archive › dc_04a6feb823aa
DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training — arXiv:2609.04094
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.04094
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T01:11:32.532Z
- SHA-256 of the archived text
ca913df935d4a7faafd9c5efaf5d5e3c031da03f1b335b9013aba944f3273a57- Stable URL of this page
https://metraipolis.com/documento/dc_04a6feb823aa
How to cite
DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training — arXiv:2609.04094 — arXiv (en), https://arxiv.org/abs/2609.04094 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_04a6feb823aa (Metraipolis archive dc_04a6feb823aa, sha256 ca913df935d4a7faafd9c5efaf5d5e3c031da03f1b335b9013aba944f3273a57)
BibTeX
@misc{dc04a6feb823aa,
title = {DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training — arXiv:2609.04094},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.04094},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_04a6feb823aa; sha256 ca913df935d4a7faafd9c5efaf5d5e3c031da03f1b335b9013aba944f3273a57},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.04094. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement Learning from Verifiable Rewards works well when a task has a programmatic checker, but most long-horizon agent domains have none. We work in the outcome-blind setting, where ground-truth success signals are not available. Multi-criteria rubrics are a popular way to supply such a reward; they are scored once per trajectory, but a single scalar is a poor signal across tens of steps. We propose DRACO: Distributing Rubric-based Advantage for Credit Optimization. It generates rubrics dynamically during training to track the policy's evolving capability, scores those rubrics once per completed trajectory, and redistributes that judgment over the steps responsible for annotated rubrics to produce differentiated per-step advantages in GRPO. The redistribution is closed-form and does not introduce any trained attribution module. On AppWorld, DRACO gains 15.9 points over the base model and 5.3 points over GRPO trained with a sparse ground-truth reward, despite not using any verifiers itself. On out-of-domain Tau-Bench, it gains 5.3 points over the base model even without a frontier judge, beating both ground-truth-reward training and other rubric-based training settings. The code for DRACO is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2609.04094 — 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 2026-10-17T00:28:02.907Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).