Metraipolis › archive › dc_67ceadbb827f
State of Thought Enables Endogenous Reasoning — arXiv:2609.16055
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.16055
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T18:32:32.333Z
- SHA-256 of the archived text
d81d5bba3edca1b2c68fc7eee45cc6ecb4f90fec86691747fd50b97bf3bbf819- Stable URL of this page
https://metraipolis.com/documento/dc_67ceadbb827f
How to cite
State of Thought Enables Endogenous Reasoning — arXiv:2609.16055 — arXiv (en), https://arxiv.org/abs/2609.16055 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_67ceadbb827f (Metraipolis archive dc_67ceadbb827f, sha256 d81d5bba3edca1b2c68fc7eee45cc6ecb4f90fec86691747fd50b97bf3bbf819)
BibTeX
@misc{dc67ceadbb827f,
title = {State of Thought Enables Endogenous Reasoning — arXiv:2609.16055},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.16055},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_67ceadbb827f; sha256 d81d5bba3edca1b2c68fc7eee45cc6ecb4f90fec86691747fd50b97bf3bbf819},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.16055. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Test-time compute has emerged as a major approach to improving the capabilities of Large Language Models (LLMs). However, existing test-time reasoning paradigms rely heavily on externally imposed control, either through fixed reasoning programs or through costly expansion in constrained search spaces, limiting both generalization and efficiency. We propose State of Thought (SoT), a new reasoning paradigm that enables endogenous reasoning in LLMs, with the model's internal reasoning state governing how reasoning unfolds. Concretely, SoT extracts a compact dynamics-geometric state from the model's internal information transfer and uses a 582-parameter controller on frozen backbones to selectively activate historical reasoning support useful under the current reasoning state, framing reasoning as a state-conditioned process over evidence rather than an externally prescribed token chain. Across quantitative (1.34x), general (1.62x), symbolic-and-code (1.76x), and long-context (2.51x) reasoning on 3 LLMs and 16 datasets, SoT consistently improves mean-baseline accuracy while reducing generated tokens by 62.6% and end-to-end latency by 44.6%. Across 2 VLM scales and 3 reasoning tasks, it improves mean accuracy by 3.8 points over reasoning baselines, with 74.9% fewer completion tokens and 73.5% lower latency than search-based methods. Under constrained access, SoT retains 38.2%/36.5% mean accuracy gains in training-free/embedding-only settings, while trajectory-only judging reaches 84.1% agreement across 3 API models. Together, endogenous state-driven reasoning provides a generalizable and efficient alternative. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-16 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.16055 — 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 2026-10-15T18:03:09.552Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).