Newest papers

SOMA: Continuous Monitoring and Intervention for System Two Reasoning with System One Models

Ruizhi Qian, Lei Shi, Tianyu Hu, Bo Yan, Song Wang, Yu Tian · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26793 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Pada manusia, metakognisi memisahkan pemantauan dari kontrol: sinyal cepat, intuitif menjaga pemantauan sementara lambat, deliberate reasoning mengembangkan dan membantu memutuskan apakah untuk melanjutkan, mengubah jalur, atau berhenti. Sistem One (S1) keputusan model, seperti Jev, adalah kandidat alami untuk peran ini: mereka mengembalikan kemungkinan terkalibrasi untuk bahasa alami pertanyaan tentang teks dalam satu panggilan cepat, namun mereka hanya digunakan sebagai gerbang sebelum alasan. Kami memperkenalkan SOMA (System One Metakognitive Awareness), di mana model S1 terus memonitor sistem Dua (S2) alasan model dan campur tangan ketika diperlukan. SOMA memiliki tiga lapisan: satu lapisan skenario menyatakan masing-masing tujuan dalam bahasa alami dengan tindakan predefined, satu lapisan kesadaran memungkinkan S1 menilai apakah perilaku yang dijelaskan meningkat, dan pada lapisan kontrol yang berbeda

Machine translation (Meta M2M100); the English original follows.

Abstract

Reasoning models often run into problems without noticing them, even when symptoms are visible in their own text. In humans, metacognition separates monitoring from control: fast, intuitive signals keep watch while slow, deliberate reasoning unfolds and help decide whether to continue, change course, or stop. System One (S1) decision models, such as Jev, are a natural candidate for this role: they return a calibrated probability for a natural-language question about a text in one fast call, yet they are used only as gates before reasoning. We introduce SOMA (System One Metacognitive Awareness), in which an S1 model continuously monitors a System Two (S2) reasoning model and intervenes when needed. SOMA has three layers: a scenario layer states each objective in natural language with predefined actions, an awareness layer lets S1 judge whether the described behavior appears, and a control layer accumulates these judgments, first reminds S2, and takes over only if the behavior persists. Sharing one S1 model and one controller across tasks, without training, SOMA addresses two different failures on two reasoning models: on overthinking, it cuts reasoning tokens by up to 36\% while improving average accuracy, and on unanswerable questions, it raises abstention from 70.5\% to 87.2\%.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Metacognition, Soma, Control (management), Natural (archaeology), Artificial intelligence

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange