Metraipolis › archive › dc_7595327ee094

RLiD: Reinforcement Learning in the Decision Plane for Language Agents and LLMs — arXiv:2611.02455

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02455
License (as archived)
CC0-1.0
Archived
2026-11-03T10:53:30.815Z
SHA-256 of the archived text
7ff45d84ccdbaa769ceb0429c513bee8692a33e25ffa2df2a4de0f4ec6e82ff6
Stable URL of this page
https://metraipolis.com/documento/dc_7595327ee094

How to cite

RLiD: Reinforcement Learning in the Decision Plane for Language Agents and LLMs — arXiv:2611.02455 — arXiv (en), https://arxiv.org/abs/2611.02455 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7595327ee094 (Metraipolis archive dc_7595327ee094, sha256 7ff45d84ccdbaa769ceb0429c513bee8692a33e25ffa2df2a4de0f4ec6e82ff6)
BibTeX
@misc{dc7595327ee094,
  title = {RLiD: Reinforcement Learning in the Decision Plane for Language Agents and LLMs — arXiv:2611.02455},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02455},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7595327ee094; sha256 7ff45d84ccdbaa769ceb0429c513bee8692a33e25ffa2df2a4de0f4ec6e82ff6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02455. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reinforcement learning (RL) has improved the capabilities of large language models (LLMs) across a wide range of tasks. However, token-level RL assigns an action's advantage to the entire generated response, including tokens specifying the parsed action, reasoning chains, and malformed content. Although the environment executes only the parsed action, variation in how the action is expressed in natural language can introduce additional variance into policy-gradient estimates. We propose RLiD (Reinforcement Learning in the Decision plane), a framework that directly optimizes the probabilities of executable action options. RLiD compiles observations, an optional compact memory, and a labeled action menu into the model input, and uses a large decision model (LDM) to select an action in a single forward pass. By eliminating autoregressive token generation and bypassing action parsing, RLiD avoids invalid action outputs and improves RL rollout efficiency. Our analysis shows that, within a fixed text policy, marginalizing over alternative realizations of the same action removes a zero-mean gradient component without changing the expected update. We further distill trained LDM policies into generative LLMs through supervision on their action trajectories. Experiments on ARC, Countdown, and Sokoban demonstrate improved performance or lower computational cost across several settings. On ALFWorld, the complete training and distillation pipeline reaches 99.3% success in 12.7 H100 GPU-hours, compared with 86.6% in 17.3 GPU-hours for token-level RL of the same LLM. With compiled memory, our strongest Qwen2.5-1.5B LDM and distilled generative LLM achieve 100% success on all 134 unseen and 140 seen ALFWorld evaluation games.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.02455 — 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 2026-11-03T10:49:38.349Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).