Metraipolis › archive › dc_31428dacdbe7

Divergent Paths: Learning an Outcome-Predictive Representation of LLM Agent Trajectories — arXiv:2610.19614

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19614
License (as archived)
CC0-1.0
Archived
2026-10-15T14:17:39.154Z
SHA-256 of the archived text
f12dfdc0ddbfb7c4be85841acc8f6c27954dfa80fe4fede19bdbb0911caba843
Stable URL of this page
https://metraipolis.com/documento/dc_31428dacdbe7

How to cite

Divergent Paths: Learning an Outcome-Predictive Representation of LLM Agent Trajectories — arXiv:2610.19614 — arXiv (en), https://arxiv.org/abs/2610.19614 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_31428dacdbe7 (Metraipolis archive dc_31428dacdbe7, sha256 f12dfdc0ddbfb7c4be85841acc8f6c27954dfa80fe4fede19bdbb0911caba843)
BibTeX
@misc{dc31428dacdbe7,
  title = {Divergent Paths: Learning an Outcome-Predictive Representation of LLM Agent Trajectories — arXiv:2610.19614},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19614},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_31428dacdbe7; sha256 f12dfdc0ddbfb7c4be85841acc8f6c27954dfa80fe4fede19bdbb0911caba843},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19614. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In this work, we reframe agentic trajectories as paths through a shared embedding space of actions and environmental observations. We show that successful and failed rollouts of the same task separate in this space, with simple per-task discriminators predicting held-out rollout outcomes at up to 0.95 AUROC across three agent benchmarks and three policy models. To exploit this signal across tasks, we introduce Precog-RM, a transformer-based model that meta-learns a representation in which outcome prediction can be performed in a few-shot setting. We show that Precog-RM generalizes to unseen tasks and models, and gives confident predictions of trajectory success long before termination. Deployed as an early stopping controller for Group Relative Policy Optimization (GRPO) on Tau2-Bench, Precog-RM cuts training FLOPs by up to 47.3\% while maintaining performance comparable to baseline GRPO.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.19614 — 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 2026-10-15T13:53:35.622Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).