Role play with large language models

2023/11/08 by Murray Shanahan, Kyle McDonell, Laria Reynolds · 1 voice · 70 citations
Computer Science · Psychology · #Speech and dialogue systems #Social Robot Interaction and HRI #AI in Service Interactions

paper · pdf · doi:10.1038/s41586-023-06647-8

Citations

• On the Dangers of Stochastic Parrots

Cited by

• Emergent Learner Agency in Implicit Human-AI Collaboration: How Supportive and Contrarian AI Personas Reshape Interaction
• Perceived AGI: Believability as Dimensional Completeness, Not Capability
• AI Fiction in the Wild
• Machine understanding
• Anthropomorphic Behaviors of AI
• "AI Psychosis" in Context: How Conversation History Shapes LLM Responses to Delusional Beliefs
• Linear representations in language models can change dramatically over a conversation
• A Pragmatic View of AI Personhood
• Technological folie à deux: Feedback Loops Between AI Chatbots and Mental Illness
• Does It Make Sense to Speak of Introspection in Large Language Models?
• LLM Social Simulations Are a Promising Research Method
• Palatable Conceptions of Disembodied Being: Terra Incognita in the Space of Possible Minds
• Identifying features that shape perceived consciousness in LLM-based AI: A quantitative study of human responses
• Why human-AI relationships need socioaffective alignment
• The Sorrows of Young Chatbot Users: Harm and Responsibility in Human-AI Relationships
• ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessions
• LLM Agents as VC investors: Predicting Startup Success via RolePlay-Based Collective Simulation
• Dual-Margin Embedding for Fine-Grained Long-Tailed Plant Taxonomy
• LLMs on Drugs: Language Models Are Few-Shot Consumers
• Fine-tuning of lightweight large language models for sentiment classification on heterogeneous financial textual data
• Adapting Like Humans: A Metacognitive Agent with Test-time Reasoning
• AI Consciousness and Existential Risk
• On the Creativity of AI Agents
• When Machines Join the Moral Circle: The Persona Effect of Generative AI Agents in Collaborative Reasoning
• Using language models to label clusters of scientific documents
• Open Character Training: Shaping the Persona of AI Assistants through Constitutional AI
• TwinVoice: A Multi-dimensional Benchmark Towards Digital Twins via LLM Persona Simulation
• Cognitive Convergence: Deep Similarities Between Large Language Models and Human Cognition
• Simulating Subjects: The Promise and Peril of Artificial Intelligence Stand-Ins for Social Agents and Interactions
• Simulacra as conscious exotica
• Ethically enslaving AI
• The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment
• Context Structure Reshapes the Representational Geometry of Language Models
• RAVR: Reference-Answer-guided Variational Reasoning for Large Language Models
• Dissecting Role Cognition in Medical LLMs via Neuronal Ablation
• Politically Speaking: LLMs on Changing International Affairs
• Enhancing Vision-Language Models for Autonomous Driving through Task-Specific Prompting and Spatial Reasoning
• Towards AI as Colleagues: Multi-Agent System Improves Structured Professional Ideation
• SMART 2.0 Statistical Metabolomics Analysis: An R Tool 2.0
• Using AI to DIY? Incorporating AI-Generated Learning Tools in the Classroom
• Individualized Cognitive Simulation in Large Language Models: Evaluating Different Cognitive Representation Methods
• Presenting Large Language Models as Companions Affects What Mental Capacities People Attribute to Them
• Prompt Optimization via Retrieved Reasoning Assets and Multi-Agent Analysis
• A Two-Step, Multidimensional Account of Deception in Language Models
• ROBOPSY PL[AI]: Using Role-Play to Investigate how LLMs Present Collective Memory
• Quantifying uncert-AI-nty: Testing the accuracy of LLMs’ confidence judgments
• FURINA: A Fully Customizable Role-Playing Benchmark via Scalable Multi-Agent Collaboration Pipeline
• Incoherence in Goal-Conditioned Autoregressive Models
• The fragility of "cultural tendencies" in LLMs
• Agentic Misalignment: How LLMs Could Be Insider Threats
• NeuroBridge: Using Generative AI to Bridge Cross-neurotype Communication Differences through Neurotypical Perspective-taking
• When Can AI Models Explain Learning? Validity Criteria for AI as Cognitive Models in Education
• Semantic-Aware Fuzzing: An Empirical Framework for LLM-Guided, Reasoning-Driven Input Mutation
• A large-scale evaluation of commonsense knowledge in humans and large language models
• Large language model-driven agents in nursing practice: A scoping review
• A Theory of Appropriateness That Accounts for Norms of Rationality
• Through the Lens of Human-Human Collaboration: A Configurable Research Platform for Exploring Human-Agent Collaboration
• Charting trajectories of human thought using large language models
• On the creativity of large language models
• Is the `Agent' Paradigm a Limiting Framework for Next-Generation Intelligent Systems?
• LLM in the Middle: A Systematic Review of Threats and Mitigations to Real-World LLM-based Systems
• We Need a New Ethics for a World of AI Agents
• Benchmark of stylistic variation in LLM-generated texts
• Getting In Contract with Large Language Models -- An Agency Theory Perspective On Large Language Model Alignment
• Probing the Preferences of a Language Model: Integrating Verbal and Behavioral Tests of AI Welfare
• Drivers of generative AI adoption in higher education through the lens of the Theory of Planned Behaviour
• The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs
• CS-Agent: LLM-based Community Search via Dual-agent Collaboration
• The benefits and dangers of anthropomorphic conversational agents
• Mockingbird: How does LLM perform in general machine learning tasks?
• Out-of-Context Abduction: LLMs Make Inferences About Procedural Data Leveraging Declarative Facts in Earlier Training Data

Discussions

• Rather than constituting a good theory of the human mind, Centaur is better characterised as good at role playing. doi.org/10.1038/s415... Based on prompt, Centaur can make human-like or super-human r [bsky, 17 points, 1 comments]