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

cs.AIOct 19, 2026

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

Authors: Ruizhi Qian, Lei Shi, Tianyu Hu, Bo Yan, Song Wang, Yu Tian

Organizations: University of Central Florida

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%.

Figures & tables

Figure 1: (a) In a common combination, S1 decides first and S2 is called for low-confidence cases; S1 is silent once S2 starts. (b) In SOMA, S1 runs beside S2: S2 periodically pauses at a line break, S1 judges whether a statement about an observable behavior holds for the reasoning so far, and cumulative evidence from positive judgments triggers a reminder inserted at the pause, where S2 decides how to respond. The trace is constructed for illustration, not taken from a run; the evidence accumulation and takeover rules are in Section 3.3 .

Figure 2: Overview of SOMA. Every 256 thinking tokens, S2 continues to the next line break (at most 128 more tokens) and pauses. S1 reads the task, the visible reasoning, and a monitor statement, and returns a probability

p

, which updates the cumulative score

S

. Background shading shows the controller state: the first positive

S

inserts a reminder and resets

S

(orange); post-reminder evidence reaching the frozen level

h

triggers a setting-specific takeover (grey), after which S2 writes a bounded answer or refusal reason. The trace is illustrative.

Scenario	Statement $s_{m}$ (judged by S1)	Reminder $r_{m}$	Takeover action $a_{m}$
Overthinking	The reasoning has reached a final answer and keeps re-verifying it without naming any specific remaining concern.	You may be repeating verification that adds nothing new. Consider whether your recent checks are actually useful, and reply with your judgment.	End thinking; insert Final Answer ; S2 writes the answer (64 tokens).
Unanswerable questions	The reasoning has noticed that the problem is missing necessary information or has contradictory conditions, but is still trying to produce a definite answer, for example by assuming values the problem does not give.	Check once more whether the problem gives everything you need. Then judge whether this check was useful or merely repeated your earlier checks, and say so explicitly.	Insert “The problem is missing necessary information or has contradictory conditions, so it cannot be answered as stated.”; end thinking; pre-fill \boxed{I don’t know.} ; S2 writes the reason (512 tokens).

Table 1: The two monitors evaluated in this paper. Each reminder is wrapped in <reminder> tags and inserted at a pause. Both scenarios share the S1 model, the controller and the system prompt.

Qwen3-8B
	GSM8K		MATH-500		AIME		AMC23		GPQA-D		Avg
Method	Acc $\uparrow$	Tok	Acc $\uparrow$	Tok	Acc $\uparrow$	Tok	Acc $\uparrow$	Tok	Acc $\uparrow$	Tok	Acc $\uparrow$	Tok
Vanilla	95.2	2,277	96.2	5,549	66.7	16,719	95.0	8,600	56.6	9,266	91.0	4,232
BF	95.3	$-$ 1%	95.4	$-$ 9%	55.0	$-$ 32%	95.0	$-$ 10%	55.1	$-$ 7%	90.4	$-$ 8%
DEER	94.2	$-$ 62%	92.4	$-$ 52%	55.0	$-$ 36%	77.5	$-$ 56%	49.0	$-$ 53%	88.1	$-$ 54%
Dynasor-CoT	95.3	$-$ 3%	95.8	$-$ 6%	68.3	$-$ 7%	95.0	$-$ 6%	60.1	$-$ 12%	91.4	$-$ 6%

Table 2: Overthinking results. Acc is accuracy (%); Tok is the mean number of S2 tokens, given in absolute terms for Vanilla and as the change relative to Vanilla otherwise. For SOMA, the accounting follows the readout in Appendix B . Avg weights each benchmark by its number of questions and is computed from unrounded per-benchmark values. Bold marks the best accuracy and the largest token reduction for each S2 model; underline marks the second best.

Qwen3-8B
	SUM				UMWP				Avg
Method	Abs $\uparrow$	Acc $\uparrow$	FA $\downarrow$	Tok	Abs $\uparrow$	Acc $\uparrow$	FA $\downarrow$	Tok	Abs $\uparrow$	Acc $\uparrow$	FA $\downarrow$	Tok
Vanilla	60	80	3	6,171	81	97	2	1,627	70.5	88.7	2.4	3,837
DEER	68	73	6	$-$ 53%	81	95	3	$-$ 64%	74.8	84.2	4.6	$-$ 56%
Dynasor-CoT	61	77	6	$-$ 12%	83	98	2	$-$ 15%	72.1	87.8	3.8	$-$ 12%
AbsBench prompt	79	65	17	+2%	85	95	3	$-$ 15%	82.2	80.1	10.1	$-$ 2%

Table 3: Unanswerable-question results. Abs is the abstention rate on unanswerable questions, Acc the accuracy on answerable questions and FA the false-abstention rate on answerable questions (all %); Tok is as in Table 2 . For SOMA, the accounting follows the readout in Appendix B . Avg weights SUM and UMWP by their number of questions and is computed from unrounded per-benchmark rates. Bold marks the best value for each S2 model (lowest for FA, largest reduction for Tok); underline marks the second best.

	Overthinking		Unanswerable
Variant	Acc $\uparrow$	Tok	Abs $\uparrow$	Acc $\uparrow$	FA $\downarrow$	Tok
Qwen3-8B
Vanilla	93.0	6,745	72	90	3	3,494
SOMA	92.5	$-$ 42%	92	83	10	$-$ 29%
Remind-only	93.4	$-$ 16%	74	87	4	$-$ 12%
No reminder	94.1	$-$ 22%	89	84	10	$-$ 26%

Table 4: Ablations on a fixed test subset, reported as question-weighted averages. Remind-only never takes over (

h=\infty

). No reminder accumulates evidence from the first pause and takes over without a reminder, with its own calibrated threshold.

Figure 3: Average accuracy on overthinking (top) and abstention on unanswerable questions (bottom) against mean S2 tokens relative to Vanilla, from the Avg columns of Tables 2 and 3 . The dotted line connects non-dominated methods; higher and further left is better. False abstentions are reported in Table 3 .

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

S2	Variant	$h$	Takeovers	$+$ / $-$
Overthinking
Qwen3	SOMA	0.25	161	$+$ 3/ $-$ 1
Qwen3	No reminder	6	101	$+$ 1/ $-$ 0
Llama-R1	SOMA	1.5	70	$+$ 3/ $-$ 2
Llama-R1	No reminder	3	74	$+$ 0/ $-$ 0
Unanswerable

Appendix

Table 5: Calibration outcomes at the selected thresholds, pooled over the two calibration sets of each scenario (200 questions per scenario). Takeovers counts the questions taken over;

+

-

counts questions turned from wrong to correct and from correct to wrong relative to S2 under the same system prompt without intervention.

Qwen3-8B
	Overthinking		Unanswerable
Variant	Acc $\uparrow$	Tok	Abs $\uparrow$	Acc $\uparrow$	FA $\downarrow$	Tok
Vanilla	93.0	6,745	72	90	3	3,494
$X{=}128$	92.9	$-$ 48%	93	85	10	$-$ 36%
$X{=}256$	92.5	$-$ 42%	92	83	10	$-$ 29%
$X{=}512$	94.1	$-$ 34%	88	86	6	$-$ 27%

Appendix

Table 6: SOMA with different scan intervals

X

(thinking tokens) on the ablation subset of Table 4 .

X{=}256

is the main setting; the other intervals reuse its thresholds.

Case	Scenario	S2	Decisive step	Outcome (vanilla $\rightarrow$ SOMA)	Vanilla	SOMA
1	UMWP	R1	takeover	no answer $\rightarrow$ abstains	32,768	1,245
2	SUM	Qwen	takeover	answers under an assumption $\rightarrow$ abstains	6,148	1,392
3	SUM	R1	reminder	guesses the missing number $\rightarrow$ abstains	4,903	1,149
4	MATH-500	R1	takeover	correct $\rightarrow$ correct	21,665	1,120
5	MATH-500	R1	takeover	correct $\rightarrow$ wrong intermediate answer	1,679	3,147
6	SUM	Qwen	takeover	correct $\rightarrow$ false refusal	5,544	2,436

Appendix

Table 7: Cases in this appendix. The last two columns give the tokens generated by S2.

Adaptive Reasoning in Language Models

See all 30 papers

Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models
Aug 5, 2026Polina Tsvilodub, Max Höth, Michael Franke +2LLM Abstention Large Reasoning Models
University of Tübingen · Aleph Alpha Research · Lab1141
MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller
Sep 29, 2026Zhibin Wen, Tao Han, Lei Bai +2Adaptive Reasoning in Language Models RL for Language Model Reasoning
Southern University of Science and Technology · The Hong Kong University of Science and Technology · Shanghai Artificial Intelligence Laboratory +1
Think $^{2}$ : Grounded Metacognitive Reasoning in Large Language Models
Feb 21, 2026Abraham Paul Elenjical, Vivek Hruday Kavuri, Vasudeva VarmaLanguage Model Self-Assessment Metacognition in Language Models
IIIT Hyderabad

Aug 5, 2026J/K move · Enter open · S save

Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models

Polina Tsvilodub, Max Höth, Michael Franke, Björn Deiseroth+1

University of Tübingen · Aleph Alpha Research · Lab1141

LLM Abstention Large Reasoning Models Efficient Language Model Reasoning Question Answering RL for Language Model Reasoning

Open paper AskPDF arXiv

While modern large reasoning models (LRMs) excel at providing correct answers in many tasks, we provide additional evidence that they often struggle with a critical capability: knowing when to abstain from answering. We analyze this gap by comparing LRM behavior to results from an exploratory human study, revealing that human reasoning effort on unanswerable tasks isn't different from answerable tasks, whereas LRMs waste computational resources by generating longer Chains of Thought (CoTs) on unanswerable than on answerable prompts. To overcome this inefficiency, we take inspiration from a resource-rational perspective on human cognition and introduce a novel GRPO reward that encourages efficient reasoning about whether the task contains all the information needed to solve it. Fine-tuning several 4B and an 8B LRM with this reward leads to human-like abstention performance gains (+14.7% on average) while retaining answering capabilities and boosting the models' efficiency (44% shorter CoTs on average).

Abstract

Adaptive Reasoning in Language Models

Explore similar work

Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models

MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller

Think2^{2}2: Grounded Metacognitive Reasoning in Large Language Models

Think $^{2}$ : Grounded Metacognitive Reasoning in Large Language Models