Eeyore: Realistic Depression Simulation via Supervised and Preference Optimization
On the map: LLM alignment / RLHF / preference optimisation → Medical LLM alignment
Abstract (arXiv)
Large Language Models (LLMs) have been previously explored for mental healthcare training and therapy client simulation, but they still fall short in authentically capturing diverse client traits and psychological conditions. We introduce Eeyore, an 8B model optimized for realistic depression simulation through a structured alignment framework, incorporating expert input at every stage. First, we systematically curate real-world depression-related conversations, extracting depressive traits to guide data filtering and psychological profile construction, and use this dataset to instruction-tune Eeyore for profile adherence. Next, to further enhance realism, Eeyore undergoes iterative preference optimization -- first leveraging model-generated preferences and then calibrating with a small set of expert-annotated preferences. Throughout the entire pipeline, we actively collaborate with domain experts, developing interactive interfaces to validate trait extraction and iteratively refine structured psychological profiles for clinically meaningful role-play customization. Despite its smaller model size, the Eeyore depression simulation outperforms GPT-4o with SOTA prompting strategies, both in linguistic authenticity and profile adherence.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Medical LLM alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 12 of its 34 references that are in the map. First posted on arXiv 21 Feb 2025.
Builds on
12 of 34 references.
Closest ancestors · 1
Same direction, newest first
Other references · 11
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
- 2017Deep reinforcement learning from human preferencescited by 3,912
- 2022Out of One, Many: Using Language Models to Simulate Human Samplescited by 752
- 2021Towards Emotional Support Dialog Systemscited by 299
- 2024Strengthening Multimodal Large Language Model with Bootstrapped Preference Optimizationcited by 75
- 2024Roleplay-doh: Enabling Domain-Experts to Create LLM-simulated Patients via Eliciting and Adhering to Principlescited by 73
- 2024The Ethics of ChatGPT in Medicine and Healthcare: A Systematic Review on Large Language Models (LLMs)cited by 45
- 2024Interactive Agents: Simulating Counselor-Client Psychological Counseling via Role-Playing LLM-to-LLM Interactionscited by 43
- 2024Step-Controlled DPO: Leveraging Stepwise Error for Enhanced Mathematical Reasoningcited by 35
- 2021Speaker and Time-aware Joint Contextual Learning for Dialogue-act Classification in Counselling Conversationscited by 31
- 2024Towards a Client-Centered Assessment of LLM Therapists by Client Simulationcited by 24
Built on by
Citing papers per year: 2025: 2; 2026: 1.
3 papers.
Similar, not cited
Close in wording, no citation either way
- 2024MentalArena: Self-play Training of Language Models for Diagnosis and Treatment of Mental Health Disorders
- 2025DepressLLM: Interpretable domain-adapted language model for depression detection from real-world narratives
- 2023Mental-LLM: Leveraging Large Language Models for Mental Health Prediction via Online Text Data
- 2026Multi-Objective Alignment of Language Models for Personalized Psychotherapy
- 2026TherapyGym: Evaluating and Aligning Clinical Fidelity and Safety in Therapy Chatbots
- 2025MindEval: Benchmarking Language Models on Multi-turn Mental Health Support
- 2024Synthetic Data Generation with LLM for Improved Depression Prediction
- 2026Dep-LLM: Training-Free Depression Diagnosis via Evidence-Guided Structured Multi-factor with Reliable LLM Reasoning