Eeyore: Realistic Depression Simulation via Supervised and Preference Optimization

ACL ARR 2025 February Submission4316 Authors

Abstract:

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 \textbf{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.

Paper Type: Long

Research Area: Human-Centered NLP

Research Area Keywords: patient simulation; expert-in-the-loop; NLP for mental health; participatory/community-based NLP;

Contribution Types: NLP engineering experiment, Publicly available software and/or pre-trained models, Data resources

Languages Studied: english

Submission Number: 4316