Metraipolis › archive › dc_0ef70fc0e5b3

State2State: Environment-Derived Mid-Training for LLM Agents — arXiv:2608.04934

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.04934
License (as archived)
CC0-1.0
Archived
2026-10-31T11:25:19.055Z
SHA-256 of the archived text
9c6db02fa0e328f44ff39e6238ca70eb268abaaf0290d06860a49bdf8682d9e4
Stable URL of this page
https://metraipolis.com/documento/dc_0ef70fc0e5b3

How to cite

State2State: Environment-Derived Mid-Training for LLM Agents — arXiv:2608.04934 — arXiv (en), https://arxiv.org/abs/2608.04934 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0ef70fc0e5b3 (Metraipolis archive dc_0ef70fc0e5b3, sha256 9c6db02fa0e328f44ff39e6238ca70eb268abaaf0290d06860a49bdf8682d9e4)
BibTeX
@misc{dc0ef70fc0e5b3,
  title = {State2State: Environment-Derived Mid-Training for LLM Agents — arXiv:2608.04934},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.04934},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0ef70fc0e5b3; sha256 9c6db02fa0e328f44ff39e6238ca70eb268abaaf0290d06860a49bdf8682d9e4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04934. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers. Though effective, both remain bottlenecked by externally specified tasks and supervision signals, limiting the scalability and diversity of agent training. We study an environment learning paradigm in which agents acquire interaction and manipulation capabilities solely through environment interaction, without externally specified tasks. We propose State2State, an environment-derived mid-training method that converts explored environment states into training objectives, challenging agents to reach a specified target state. By deriving tasks from environment exploration and verifying success through rule-based state matching, State2State provides scalable and verifiable training objectives without expert supervision or manual task design. Experiments on ALFWorld and ScienceWorld show that State2State improves agent performance as a standalone environment-learning stage in most settings. As initialization for downstream RL, it further improves final performance and learning efficiency, with promising evidence of cross-environment generalization.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-06 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.04934 — 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 2026-10-31T11:13:23.193Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).