EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Dynamic Closed-Loop Steering for Adaptive and Geometry-Preserving System-2 Reasoning

Deyu Meng, Tongchuan Xia, Yuanxin Cai

Applied Sciences · 27.06.2026 · 16 (13) · 6423

DOI: 10.3390/app16136423 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 27.06.2026. Verfasst von Deyu Meng und 2 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Beijing Municipal High-Level Overseas Talent Funding.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language models (LLMs) are undergoing a paradigm shift from fast generation to deliberate System-2 reasoning, where scaling test-time compute is critical for unlocking complex reasoning capabilities. Yet, more computation does not guarantee better reasoning: under unconstrained test-time scaling, models frequently fall into “overthinking” traps, repeatedly elaborating on incorrect trajectories while wasting token budgets. We propose Dynamic Closed-Loop Steering, a training-free framework that treats reasoning as a monitored control process rather than a static decoding run. The framework follows a sense–decide–act design: lightweight latent signals sense trajectory divergence, a feedback regulator determines adaptive intervention timing and intensity, and a geometry-preserving actuation step redirects hidden states while maintaining their norm structure. This design emphasizes adaptive control and macro-level trajectory observability over static hidden-state addition. Across mathematical and reasoning benchmarks, Dynamic Closed-Loop Steering improves final-answer accuracy while curbing excess token consumption, yielding Pareto-favorable trade-offs across the large majority of evaluated configurations, with graceful degradation near the model’s capability floor. Trajectory analysis confirms the method effectively suppresses repetitive self-justification and accelerates recovery from erroneous episodes. These results establish closed-loop, geometry-preserving intervention as a practical foundation for robust System-2 reasoning.

Förderung

  • Beijing Municipal High-Level Overseas Talent Funding

Zitieren

APA

Meng, D., Xia, T., & Cai, Y. (2026). Dynamic Closed-Loop Steering for Adaptive and Geometry-Preserving System-2 Reasoning. Applied Sciences, 16(13), 6423. https://doi.org/10.3390/app16136423

Vancouver

Meng D, Xia T, Cai Y. Dynamic Closed-Loop Steering for Adaptive and Geometry-Preserving System-2 Reasoning. Applied Sciences. 2026;16(13):6423. doi:10.3390/app16136423

BibTeX
@article{Meng2026,
  author = {Meng, Deyu and Xia, Tongchuan and Cai, Yuanxin},
  title = {Dynamic Closed-Loop Steering for Adaptive and Geometry-Preserving System-2 Reasoning},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {13},
  pages = {6423},
  doi = {10.3390/app16136423}
}
Kostenlos starten →