Confident but Conflicted: Internal Uncertainty and Cognitive Dissonance Resolution in LLMs

2026-06-21

AI Summary

This study addresses the unclear relationship between large language models’ mechanisms for resolving cognitive dissonance—when confronted with information conflicting with their prior outputs—and their internal uncertainty. By systematically manipulating source authority and evidence quality across twelve health-related claims, the work introduces “Trust Elasticity,” a novel metric inspired by econometrics, to empirically link model behavioral shifts to internal uncertainty for the first time. Experiments on Qwen and Llama model families reveal significant inter-model differences in trust elasticity, near-zero elasticity across all models when presented with clearly false claims, and a strong predictive capacity of internal uncertainty measures for behavioral responses.

Abstract

Large language models (LLMs) frequently encounter inputs that disagree with their prior outputs, through user pushback, retrieved documents, or web search results. While the way they resolve such conflicts -- a process we frame as cognitive dissonance resolution -- has been characterized behaviorally, its connection to internal model uncertainty is not well understood. To study this systematically, we vary persuasion attempts along two dimensions, source authority and evidence quality, across 12 health-science claims of stratified epistemic status. Dissonance can be resolved through persuasion, backfire, or immunity. We introduce Trust Elasticity (TE), an econometrics-inspired measure of how readily a model is persuaded toward conflicting evidence. Across four LLMs, TE varies substantially, while clearly false claims elicit near-zero TE across all models. On two open-weight models, we further find that this variation is associated with two complementary internal uncertainty indicators, Confidence Miscalibration in Qwen and Internal Uncertainty Change in Llama. These results link cross-model behavioral variation to a measurable internal property and point to interventions targeting internal uncertainty as future work.

Problem

• cognitive dissonance
• internal uncertainty
• large language models
• persuasion
• Trust Elasticity

Innovation

• Trust Elasticity
• Internal Uncertainty
• Cognitive Dissonance Resolution
• Confidence Miscalibration
• Large Language Models