Research map / Large language models / AI and society
Can LLMs Move Beyond Short Exchanges to Realistic Therapy Conversations?
Zhengqing Yuan, Liang Wu, Jian Xu, Zheyuan Zhang, Kaiwen Shi, Weixiang Sun, Lichao Sun, Yanfang Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| rephrase public exam items by clinicians | 18 LLMs · CareBench-CBT knowledge QA | accuracy: 92.31 -> 71.17 | Hurt |
| add Hierarchical Therapy Memory (HTM) (ablation) | GPT-5 · CareBench-CBT multi-turn | whole-session score: 4.10 -> 4.45 | Improved |
| add HTM (ablation) | LLaMA-3.3-70B · CareBench-CBT multi-turn | turn score (model-based history): 1.47 -> 1.85 | Improved |
| add HTM (ablation) | Qwen-2.5-7B · CareBench-CBT multi-turn | turn score (model-based history): 1.93 -> 1.97 | No clear effect |
| add HTM (ablation) | GPT-5 · CareBench-CBT multi-turn | normalized latency: 1.12x -> 0.65x | Improved |
| use human counselor history instead of model-generated history (ablation) | 18 LLMs · CareBench-CBT multi-turn | turn score: average gap -0.01 | No clear effect |
| use human-based history instead of model history (ablation) | GPT-oss-120B · CareBench-CBT multi-turn | turn score: -0.40 | Hurt |
| scale to large open-weight model on vignette classification | GPT-oss-120B · CareBench-CBT case vignettes | accuracy: 41.7% | No clear effect |
| add explicit length constraints or brevity prompts | LLMs · CareBench-CBT multi-turn | counseling quality | Hurt |
Closest papers by meaning
- TherapyGym: Evaluating and Aligning Clinical Fidelity and Safety in Therapy Chatbots (NeurIPS 2026 · similarity 0.92)
- Position: Beyond Assistance – Reimagining LLMs as Ethical and Adaptive Co-Creators in Mental Health Care (ICML 2025 · similarity 0.91)
- An LLM-based Simulation Framework for Embodied Conversational Agents in Psychological Counseling (AAAI 2026 · similarity 0.87)
- MedJourney: Benchmark and Evaluation of Large Language Models over Patient Clinical Journey (NeurIPS 2024 · similarity 0.86)
- Thousand Voices of Trauma: A Large-Scale Synthetic Dataset for Modeling Prolonged Exposure Therapy Conversations (NeurIPS 2025 · similarity 0.86)
- CareBot: A Pioneering Full-Process Open-Source Medical Language Model (AAAI 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
