Metraipolis › archive › dc_4458607ffed0

Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance — arXiv:2609.24066

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.24066
License (as archived)
CC0-1.0
Archived
2026-10-09T06:50:02.306Z
SHA-256 of the archived text
62c872c7dc99c5dbe82bebe747cb018171aaf52a5d7866f56c663e1dae3155ae
Stable URL of this page
https://metraipolis.com/documento/dc_4458607ffed0

How to cite

Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance — arXiv:2609.24066 — arXiv (en), https://arxiv.org/abs/2609.24066 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4458607ffed0 (Metraipolis archive dc_4458607ffed0, sha256 62c872c7dc99c5dbe82bebe747cb018171aaf52a5d7866f56c663e1dae3155ae)
BibTeX
@misc{dc4458607ffed0,
  title = {Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance — arXiv:2609.24066},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.24066},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4458607ffed0; sha256 62c872c7dc99c5dbe82bebe747cb018171aaf52a5d7866f56c663e1dae3155ae},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24066. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Best-of-$N$ is a widely used inference strategy for complex reasoning, whose effectiveness depends on whether sampled candidates can cover diverse and high-quality reasoning paths. However, post-trained reasoning models often suffer from \emph{exploration collapse}, where independent rollouts repeatedly follow similar reasoning paths and limit the gains from increasing the rollout budget. Existing methods alleviate this issue by promoting broader exploration, but do not explicitly guide exploration toward continuations that make meaningful progress, resulting in limited exploration efficiency. To address this, we propose \emph{\underline{S}tate-conditioned \underline{P}rogress-guided \underline{S}teering} (SPS), a training-free latent steering framework. Specifically, SPS constructs a state-conditioned Direction Bank containing multiple progress-guided steering vectors for different prefix-state regions. During online inference, SPS retrieves a suitable steering vector based on the current prefix state and applies it at high-uncertainty transitions to guide the next reasoning step toward meaningful progress. Extensive experiments across multiple model scales and benchmarks demonstrate that SPS consistently outperforms strong baselines. Further analyses validate the effectiveness of its key designs and offer valuable insights for future research. The code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.24066 — 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 2026-10-09T06:48:45.025Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).