Large language models detect amplified abstract interpretation in trait anxiety

Research on the relation of interpretation biases to internalizing symptoms has primarily focused on emotional valence (i.e., negativity bias), leaving other aspects of cognitive processing underexplored. Guided by dual-coding theory (Kousta et al., 2011; Paivio, 1991), we examined how the balance between abstract versus concrete elaboration shapes social interpretations in adults with varying levels of trait anxiety and depressive symptoms. A confirmatory sample of 158 U.S. participants wrote interpretations of 24 ambiguous social images (12 negative, 12 nonnegative) presented with gradually decreasing blur (Stage 1 = 80% blur, Stage 2 = 20%, Stage 3 = 0%). Large language model (Generative Pre-Trained Transformer-4.1) analyses were used to compute emotional valence and the proportion of abstract versus concrete language in each response. Trait anxiety symptoms emerged as robust correlates of language-based interpretation biases. As preregistered, both trait anxiety and depressive symptoms correlated with more negative interpretations. This pattern was stronger for negative than nonnegative scenarios, replicating previous findings of negativity bias. In addition, in exploratory analyses, individuals high in trait anxiety, but not depressive symptoms, showed an amplified increase in abstract relative to concrete language across stages. The result held even when accounting for lexicon-based abstractness measures. These results extend valence-based theories by suggesting that trait anxiety symptoms are associated with amplified abstraction during social interpretations. Large language model-derived linguistic indices provide a nuanced, automated method for capturing dual-coding dynamics in natural language, broadening avenues to probe latent cognitive processing patterns in psychopathology. (PsycInfo Database Record (c) 2026 APA, all rights reserved).