Reasoning aligns language models to human cognition

TL;DR

This paper investigates how language models reason under uncertainty compared to humans, highlighting the importance of chain-of-thought reasoning in improving inference and aligning model behavior with human cognition.

Contribution

It introduces a new probabilistic reasoning benchmark and a mechanistic model to analyze and compare human and language model decision-making processes.

Findings

Extended reasoning improves inference performance.

Chain-of-thought reasoning aligns models closer to human belief trajectories.

Active sampling shows modest gains from reasoning.

Abstract

Do language models make decisions under uncertainty like humans do, and what role does chain-of-thought (CoT) reasoning play in the underlying decision process? We introduce an active probabilistic reasoning task that cleanly separates sampling (actively acquiring evidence) from inference (integrating evidence toward a decision). Benchmarking humans and a broad set of contemporary large language models against near-optimal reference policies reveals a consistent pattern: extended reasoning is the key determinant of strong performance, driving large gains in inference and producing belief trajectories that become strikingly human-like, while yielding only modest improvements in active sampling. To explain these differences, we fit a mechanistic model that captures systematic deviations from optimal behavior via four interpretable latent variables: memory, strategy, choice bias, and occlusion awareness. This model places humans and models in a shared low-dimensional cognitive space, reproduces behavioral signatures across agents, and shows how chain-of-thought shifts language models toward human-like regimes of evidence accumulation and belief-to-choice mapping, tightening alignment in inference while leaving a persistent gap in information acquisition.