Think-Aloud Reshapes Automated Cognitive Model Discovery Beyond Behavior
For cognitive scientists and AI researchers, this work demonstrates that process-level language data can systematically reshape discovered cognitive models, addressing the under-determination problem in behavioral modeling.
The paper shows that incorporating Think-Aloud traces as additional constraints during automated cognitive model discovery improves predictive performance on held-out data and shifts model structures for 69.4% of participants from Explicit comparator to Integrated utility, enabling identification of mechanisms not recoverable from behavior alone.
Computational cognitive models discovered using large language models have so far relied solely on behavioral data. However, it is well-known that models produced from the behavioral trajectory alone are typically under-determined. In this work, we explore the use of Think Aloud traces as an additional form of data constraint during automated model discovery. When applied to the domain of risky decision-making, we find that the models discovered with think-aloud achieve significantly improved predictive performance on held-out data. Additionally, we find that the discovered models belong to different structural classes than those discovered from behavior alone for the majority of participants (69.4\%), specifically, it shifts from Explicit comparator towards Integrated utility. These results suggest that process-level language data not only improve model fit, but also systematically reshape the structure of the discovered cognitive models, enabling the identification of mechanisms that are not recoverable from behavior alone.