Metraipolis › archive › dc_0bb57394a063
Where to Start, How to Move: Analyzing Curricula under GRPO for Long-Horizon LLM Agents — arXiv:2610.19349
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.19349
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T14:15:26.456Z
- SHA-256 of the archived text
db2cd1e9dca275fe43b3b5a50ba94e77be5fb01f3dd9a5b80975c488319be42e- Stable URL of this page
https://metraipolis.com/documento/dc_0bb57394a063
How to cite
Where to Start, How to Move: Analyzing Curricula under GRPO for Long-Horizon LLM Agents — arXiv:2610.19349 — arXiv (en), https://arxiv.org/abs/2610.19349 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0bb57394a063 (Metraipolis archive dc_0bb57394a063, sha256 db2cd1e9dca275fe43b3b5a50ba94e77be5fb01f3dd9a5b80975c488319be42e)
BibTeX
@misc{dc0bb57394a063,
title = {Where to Start, How to Move: Analyzing Curricula under GRPO for Long-Horizon LLM Agents — arXiv:2610.19349},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.19349},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0bb57394a063; sha256 db2cd1e9dca275fe43b3b5a50ba94e77be5fb01f3dd9a5b80975c488319be42e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19349. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Group Relative Policy Optimization (GRPO) provides no reward-relative learning signal when all rollouts receive identical rewards. In long-horizon tasks that require multiple interactions, rewards are given only upon successful task completion, making all-fail groups common early in training and limiting opportunities to practice later steps. Guide Assisted GRPO replays a successful reference prefix before the learner continues. We study how initial positions, handoff units, and movement rules affect unassisted task completion. Across three model sizes on BFCL, starting near completion is more effective under a fixed turn-level threshold rule. Finer handoffs alone do not consistently help, but outcome-informed tool-call selection outperforms GRPO on trained and untrained task categories. Guided starts also improve AppWorld success on both evaluation splits. These findings highlight where learning starts and how starting positions move as key curriculum choices. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.19349 — 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 2026-10-15T13:53:35.576Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).