mlsanityData to 14 Oct 2026

Examining Identity Drift in Conversations of LLM Agents

arXiv 2412.00804 ↗First posted 1 Dec 2024cs.CY cs.CLcited by 25

On the map: LLM agents and tool use → Social simulation and multi-agent debate

Abstract (arXiv)

Large Language Models (LLMs) show impressive conversational abilities but sometimes show identity drift problems, where their interaction patterns or styles change over time. As the problem has not been thoroughly examined yet, this study examines identity consistency across nine LLMs. Specifically, we (1) investigate whether LLMs could maintain consistent patterns (or identity) and (2) analyze the effect of the model family, parameter sizes, and provided persona types. Our experiments involve multi-turn conversations on personal themes, analyzed in qualitative and quantitative ways. Experimental results indicate three findings. (1) Larger models experience greater identity drift. (2) Model differences exist, but their effect is not stronger than parameter sizes. (3) Assigning a persona may not help to maintain identity. We hope these three findings can help to improve persona stability in AI-driven dialogue systems, particularly in long-term conversations.

Key facts. Cited by 25 papers in the map (data through 14 Oct 2026). Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use. Builds on 15 of its 69 references that are in the map. First posted on arXiv 1 Dec 2024.

Builds on

15 of 69 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024PersonaGym: Evaluating Persona Agents and LLMs25 Jul
  2. 2023MetaAgents: Large Language Model Based Agents for Decision-Making on Teaming10 Oct
  3. 2023Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View3 Oct

Other references · 12

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2024Qwen2 Technical Reportcited by 2,120
  4. 2024Evaluating Very Long-Term Conversational Memory of LLM Agentscited by 830
  5. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 817
  6. 2022BERTopic: Neural topic modeling with a class-based TF-IDF procedurecited by 801
  7. 2024A Survey on the Memory Mechanism of Large Language Model based Agentscited by 634
  8. 2024Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalizationcited by 208
  9. 2024SOTOPIA-π: Interactive Learning of Socially Intelligent Language Agentscited by 93
  10. 2023Let the LLMs Talk: Simulating Human-to-Human Conversational QA via Zero-Shot LLM-to-LLM Interactionscited by 70
  11. 2023Learning to Memorize Entailment and Discourse Relations for Persona-Consistent Dialoguescited by 17
  12. 2024Enhancing Supermarket Robot Interaction: A Multi-Level LLM Conversational Interface for Handling Diverse Customer Intentscited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 7; 2026: 18.

25 papers.

Sort:

By area:

  1. 2025The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMscited by 64
  2. 2025Agentic AI Security: Threats, Defenses, Evaluation, and Open Challengescited by 28
  3. 2025On the Failure of Latent State Persistence in Large Language Modelscited by 9
  4. 2025Echoing: Identity Failures when LLM Agents Talk to Each Othercited by 9
  5. 2026Mitigating Conversational Inertia in Multi-Turn Agentscited by 7
  6. 2026Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systemscited by 5
  7. 2026ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessionscited by 5
  8. 2025The Narrative Continuity Test: A Conceptual Framework for Evaluating Identity Persistence in AI Systemscited by 3
  9. 2025Neural Transparency: Mechanistic Interpretability Interfaces for Anticipating Model Behaviors for Personalized AIcited by 3
  10. 2025Pay What LLM Wants: Can LLM Simulate Economics Experiment with 522 Real-human Persona?cited by 2
  11. 2026Identifying and Mitigating Bottlenecks in Role-Playing Agents: A Systematic Study of Disentangling Character Profile Axescited by 2
  12. 2026Probing the Lack of Stable Internal Beliefs in LLMscited by 2
13 more
  1. 2026MUSE: Multi-Domain Chinese User Simulation via Self-Evolving Profiles and Rubric-Guided Alignmentcited by 2
  2. 2026Belief Engine: Configurable and Inspectable Stance Dynamics in Multi-Agent LLM Deliberationcited by 2
  3. 2026Incoherent Values? Probing LLM Preferences Through Parametric Variationcited by 2
  4. 2026Decoding the Human Factor: High Fidelity Behavioral Prediction for Strategic Foresightcited by 1
  5. 2026Quantitative Introspection in Language Models: Tracking Emotive States Across Conversationcited by 1
  6. 2026FinPersona-Bench: A Benchmark for Longitudinal Psychometric Stability of Autonomous Financial Agentscited by 1
  7. 2026On the Limits of Steering Vectors for Preference-Aligned Generationcited by 1
  8. 2026Deco: Extending Cherished Physical Objects into AI Companion Agents through Dual Embodimentcited by 0
  9. 2026Measuring What Persists: Conditioning Mechanisms and a Geometric Framework for AI Agent Identitycited by 0
  10. 2026When Counterbalancing Hides the Bias: Access-Conditioned Position Lock in Forced-Choice LLM Evaluationcited by 0
  11. 2026AI YOU Town: Make Friends and Money with Your Digital Twincited by 0
  12. 2026ANIMASK: What the Model Contributes to Role Play in Simulated Story Worldscited by 0
  13. 2026The Persona Is Still There, but Who Is Speaking? Latent Identity Reversion in Persistent AI Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024LLM Agents in Interaction: Measuring Personality Consistency and Linguistic Alignment in Interacting Populations of Large Language Models
  2. 2025Persistent Instability in LLM's Personality Measurements: Effects of Scale, Reasoning, and Conversation History
  3. 2025Can LLM Agents Maintain a Persona in Discourse?
  4. 2025Consistently Simulating Human Personas with Multi-Turn Reinforcement Learning
  5. 2025Are LLM Agents Behaviorally Coherent? Latent Profiles for Social Simulation
  6. 2026Toward Robust Personalized Alignment for LLMs: Mitigating Persona Drift in Multi-Turn Dialogue
  7. 2025Agent Identity Evals: Measuring Agentic Identity
  8. 2024Building Better AI Agents: A Provocation on the Utilisation of Persona in LLM-based Conversational Agents