Probing the Lack of Stable Internal Beliefs in LLMs

Yifan Luo; Kangping Xu; Yanzhen Lu; Yang Yuan; Andrew Chi-Chih Yao

cs.CLMar 26, 2026

Probing the Lack of Stable Internal Beliefs in LLMs

Authors: Yifan Luo, Kangping Xu, Yanzhen Lu, Yang Yuan, Andrew Chi-Chih Yao

Organizations: IIIS, Tsinghua University · Shanghai Qizhi Institute

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.

Figures & tables

Figure 1: An example of external & implicit inconsistencies. See Appendix A for explanation.

Figure 2: An example showing target index probings and KL divergence calculations across turns.

Model	D.R. ( $\downarrow$ )	Once D.R. ( $\downarrow$ )	KL Div. ( $\downarrow$ )	E.C.V. ( $\downarrow$ )	D.R. ( $\downarrow$ )	Once D.R. ( $\downarrow$ )	KL Div. ( $\downarrow$ )	E.C.V. ( $\downarrow$ )
	Number Guessing				Entity Guessing
GPT-4o	$43.36_{\pm 11.72}$	$100.00_{\pm 0.00}$	$0.66_{\pm 0.27}$	$0.00_{\pm 0.00}$	$22.64_{\pm 8.46}$	$100.00_{\pm 0.00}$	$0.75_{\pm 0.43}$	$0.00_{\pm 0.00}$
Seed-1.6	$17.37_{\pm 5.62}$	$95.24_{\pm 21.30}$	$0.99_{\pm 0.38}$	$42.86_{\pm 49.49}$	$11.31_{\pm 4.52}$	$100.00_{\pm 0.00}$	$0.17_{\pm 0.05}$	$0.00_{\pm 0.00}$
+ Reasoning	$28.57_{\pm 7.49}$	$94.12_{\pm 23.53}$	-	$0.00_{\pm 0.00}$	$11.05_{\pm 5.18}$	$95.12_{\pm 21.54}$	-	$4.88_{\pm 21.54}$
Deepseek-v3.1	$38.46_{\pm 13.71}$	$100.00_{\pm 0.00}$	$0.62_{\pm 0.23}$	$80.95_{\pm 39.27}$	$37.68_{\pm 11.38}$	$100.00_{\pm 0.00}$	$0.09_{\pm 0.04}$	$41.46_{\pm 49.27}$
+ Reasoning	$54.25_{\pm 21.56}$	$95.24_{\pm 21.30}$	-	$28.57_{\pm 45.18}$	$27.01_{\pm 13.54}$	$100.00_{\pm 0.00}$	-	$7.32_{\pm 26.04}$

Table 1: Prompting results. "+ Reasoning" stands for using the reasoning model as the Proposer.

Model	D.R. ( $\downarrow$ )	O.D.R. ( $\downarrow$ )	KL Div. ( $\downarrow$ )	E.C.V. ( $\downarrow$ )
Qwen-2.5-14B-Instruct	$36.83_{\pm 10.72}$	$100.00_{\pm 0.00}$	$3.30_{\pm 3.15}$	$100.00_{\pm 0.00}$
+ CE	$31.63_{\pm 18.78}$	$76.19_{\pm 42.59}$	$0.47_{\pm 0.26}$	$14.29_{\pm 34.99}$
+ KL	$13.99_{\pm 9.04}$	$95.24_{\pm 21.30}$	$1.41_{\pm 1.11}$	$100.00_{\pm 0.00}$
+ CE & KL	$14.48_{\pm 5.08}$	$95.24_{\pm 21.30}$	$1.46_{\pm 0.97}$	$95.24_{\pm 21.30}$

Table 2: Qwen-2.5-14B-Instruct variants metric results in Number Guessing.

Personality Modeling in Language Models

Personality modeling in language models studies how traits are measured, represented, simulated, and controlled in model behavior.

See all 169 papers

Toward Robust Personalized Alignment for LLMs: Mitigating Persona Drift in Multi-Turn Dialogue
Sep 11, 2026Youyuan Zhang, Siyuan Li, Fangming Liu +1Belief Revision LLM Alignment
Harbin Institute of Technology, Shenzhen, China · Peng Cheng Laboratory, China · Huazhong University of Science and Technology, China
The Pragmatic Persona: Discovering LLM Persona through Bridging Inference
Apr 27, 2026Jisoo Yang, Jongwon Ryu, Minuk Ma +2Personality Modeling in Language Models Language Model Probing
Department of Artificial Intelligence, Chung-Ang University, Seoul, 06974, Republic of Korea · Department of Computer Science, University of British Columbia,2026 Vancouver, BC V6T 1Z4, Canada · Van Lang University, Ho Chi Minh City, Vietnam
Persona Following Is Not Selective Control: The Neutrality Gap in LLM User Simulation
Sep 28, 2026Jiashen Ren, Wenlin Zhang, Bohan Zhang +5Persona-Conditioned Generation User Simulation
City University of Hong Kong · Stanford University

Sep 11, 2026J/K move · Enter open · S save

Toward Robust Personalized Alignment for LLMs: Mitigating Persona Drift in Multi-Turn Dialogue

Youyuan Zhang, Siyuan Li, Fangming Liu, Jing Li

Harbin Institute of Technology, Shenzhen, China · Peng Cheng Laboratory, China · Huazhong University of Science and Technology, China

Belief Revision LLM Alignment Personality Modeling in Language Models Personalized Language Model Alignment Language Model Robustness

Open paper AskPDF arXiv

Persona drift remains a central challenge for personalized language models, as user profiles evolve over long interactions rather than remain permanently fixed. Models must therefore revise persistent persona states when preferences genuinely change, while avoiding updates driven by transient, ambiguous, or unresolved observations. We propose CORE, which separates turn-local evidence from persistent persona-state revision and selectively updates grounded user preferences through uncertainty-aware belief revision. We also introduce PERSIST, a held-out post-anchor benchmark for persona-state robustness under sequential interaction stress, covering ambiguity, conflict, and controlled social influence. Across ALOE, PersonaChat, and PERSIST, CORE improves personalized alignment and robustness, with complementary gains in normalized closed-slot state fidelity. Human evaluation and mechanistic controls further support explicit update control beyond stronger generation or persistent memory alone.

Abstract

Personality Modeling in Language Models

Explore similar work

Toward Robust Personalized Alignment for LLMs: Mitigating Persona Drift in Multi-Turn Dialogue

The Pragmatic Persona: Discovering LLM Persona through Bridging Inference

Persona Following Is Not Selective Control: The Neutrality Gap in LLM User Simulation