Eeyore: Realistic Depression Simulation via Supervised and Preference Optimization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916636173271040 |
|---|---|
| author | Liu, Siyang Brie, Bianca Li, Wenda Biester, Laura Lee, Andrew Pennebaker, James Mihalcea, Rada |
| author_facet | Liu, Siyang Brie, Bianca Li, Wenda Biester, Laura Lee, Andrew Pennebaker, James Mihalcea, Rada |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_00018 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Eeyore: Realistic Depression Simulation via Supervised and Preference Optimization Liu, Siyang Brie, Bianca Li, Wenda Biester, Laura Lee, Andrew Pennebaker, James Mihalcea, Rada Computation and Language Artificial Intelligence I.2.7 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. |
| title | Eeyore: Realistic Depression Simulation via Supervised and Preference Optimization |
| topic | Computation and Language Artificial Intelligence I.2.7 |
| url | https://arxiv.org/abs/2503.00018 |