Skip to content

Similar papers

Open access Sep 2026

Adaptive judgment in the cognitive reflection test: A computational analysis.

The model generates the intuitive errors elicited by the bat-and-ball problem and explains why these errors can be seen as byproducts of adaptive cognition, and shows how such judgments can be understood as rational adaptations to the learning environment.

Yi-Jin Hu, Sudeep Bhatia · 0 citations
Open access Sep 2026

When deliberation becomes inefficient.

Dual-process theories often assume that reflective, effortful thinking yields superior outcomes relative to intuitive processing. However, many decision contexts, such as chance-based or low-stakes choices, are ones in which additional deliberation cannot improve outcomes. Across two experiments (N = 918), we examined...

Brian A. Polin, Eyal Benisaac, I. Aharon · 0 citations
#natural language process... Preprint Oct 2026

The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models

Reasoning models allocate extra computation at inference time and present their answers as the product of deliberate thought. If this deliberation works the way dual-process accounts of human cognition suggest, longer thinking should weaken the classic decision biases that fast, intuitive judgment produces. Using 30 vi...

Obada Kraishan · 0 citations
#machine learning Preprint Sep 2026

Counting on Thinking: Tracing Evidence Integration in Language Models

This work asks why evidence integration in large language models requires thinking, and reasoning trajectories showed models revisiting input, recounting letters, and checking intermediate counts that informed the answer, suggesting that thinking constructs the accumulated count that direct responses lack rather than r...

Jing-Ming Xue, Robert C. Wilson, Hua-Dong Xiong · 0 citations
#artificial intelligence Preprint Sep 2026

Metacognitive Steering: Learning the Structure of Scientific Judgment

This work introduces Metacognitive Steering, an inference-time controller that reads the model's cognitive regime and dynamically composes layer-specific interventions for exploration, procedural convergence, or critical reassessment without modifying model parameters.

Vincent Karpf, Joseph Reth, Eike Gerhardt et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.