aiqbench analysis
This study challenges the assumption that human reasoning relies solely on abstract world models by demonstrating that both humans and LLMs exhibit similar pattern-matching errors in everyday causal scenarios. By isolating specific attention heads within 25 large language models, researchers show these mechanisms can predict seemingly inexplicable failures caused by irrelevant prompt details in people as well. The findings suggest a fundamental convergence between biological cognition and neural network behavior regarding how common-sense reasoning is actually processed.
Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning
arXiv:2606.13607v1 Announce Type: new Abstract: When large language models (LLMs) fail to generalize or make haphazard errors in reasoning, it is often taken as evidence that LLMs are not truly reasoning, but rather performing a kind of pattern matching. The implication is that people's behavior does not exhibit the same types of failures because human reasoning uses principled and abstract world models. We evaluate human participants and 25 LLMs on their ability to engage in common-sense reasoning about a variety of everyday situations and observe similar patterns of errors in both people and models. We then identify the set of attention heads driving LLM responses and find that these heads implement a form of pattern-matching. These attention heads allow us to predict seemingly inexplicable reasoning errors in people caused by ostensibly irrelevant prompt details. Taken together, our results suggest that everyday causal reasoning in people and LLMs is more consistent with a form of pattern-matching than with abstract world models.
Abstract:When large language models (LLMs) fail to generalize or make haphazard errors in reasoning, it is often taken as evidence that LLMs are not truly reasoning, but rather performing a kind of pattern matching. The implication is that people's behavior does not exhibit the same types of failures because human reasoning uses principled and abstract world models. We evaluate human participants and 25 LLMs on their ability to engage in common-sense reasoning about a variety of everyday situations and observe similar patterns of errors in both people and models. We then identify the set of attention heads driving LLM responses and find that these heads implement a form of pattern-matching. These attention heads allow us to predict seemingly inexplicable reasoning errors in people caused by ostensibly irrelevant prompt details. Taken together, our results suggest that everyday causal reasoning in people and LLMs is more consistent with a form of pattern-matching than with abstract world models.
Artificial Intelligence (cs.AI)
arXiv:2606.13607 [cs.AI]
(or arXiv:2606.13607v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2606.13607 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 249 words
Related coverage
Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning
arXiv · Oct 2, 2026
Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning
arXiv · Jun 16, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026
The Cost of Utopias (a Dialog)
LessWrong · Sep 18, 2026