Probing the Lack of Stable Internal Beliefs in LLMs

2026-03-26

AI Summary

This study addresses the challenge that large language models struggle to maintain unstated implicit goals across multi-turn interactions, leading to unstable simulation of personality traits such as reliability. To investigate this, the authors propose a "20 Questions"-style puzzle game paradigm in which the model secretly selects a target and responds only with “yes” or “no” to user guesses, enabling systematic evaluation of its implicit consistency. Through controlled dialog experiments, behavioral consistency metrics, and context ablation analyses, the work reveals— for the first time—that current models lack stable internal belief representations, with their implicit goals drifting significantly across conversation turns. The study introduces “implicit consistency” as a novel dimension for assessing personality modeling capabilities, highlighting a critical limitation in the models’ ability to sustain coherent objectives over extended interactions.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Reasoning with Beliefs

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Persona-driven large language models (LLMs) require consistent behavioral tendencies across interactions to simulate human-like personality traits, such as persistence or reliability. However, current LLMs often lack stable internal representations that anchor their responses over extended dialogues. This work explores whether LLMs can maintain "implicit consistency", defined as persistent adherence to an unstated goal in multi-turn interactions. We designed a 20-question-style riddle game paradigm where an LLM is tasked with secretly selecting a target and responding to users' guesses with "yes/no" answers. Through evaluations, we find that LLMs struggle to preserve latent consistency: their implicit "goals" shift across turns unless explicitly provided their selected target in context. These findings highlight critical limitations in the building of persona-driven LLMs and underscore the need for mechanisms that anchor implicit goals over time, which is a key to realistic personality modeling in interactive applications such as dialogue systems.

Problem

• stable internal beliefs
• implicit consistency
• persona-driven LLMs
• latent goal stability
• behavioral consistency

Innovation

• implicit consistency
• persona-driven LLMs
• internal belief stability
• multi-turn dialogue
• goal anchoring