Cognitive Behavioral Therapy (CBT) is a well-established, evidence-based treatment for Major Depressive Disorder. Unfortunately, there exist significant barriers to individuals accessing CBT, including cost, scarcity of therapists and stigma. This study explores the feasibility of fine-tuning small open weight large language models (LLMs) to deliver CBT for depression. Using 58 sets of synthetic CBT transcripts generated by the Nous Research fine-tune of Llama 3.1 405b, we fine-tuned three models: Mistral 7b v0.3, Qwen 2.5 7b, and Llama 3.1 8b. CBT fidelity was evaluated through a modified Cognitive Therapy Rating Scale (CTRS). All fine-tuned models were compared against each other, as well as their instruct-tuned variants. Simulated patient transcripts were generated for the purpose of evaluating model performance, with the instruct and CBT-tuned models acting as the therapist and DeepSeek-V2.5 acting as the patient. These simulated transcripts were evaluated on a modified CTRS by Gemini 1.5 Pro-002. Our findings demonstrated that the CBT-tuned models significantly outperformed their instruct-tuned counterparts, with an average improvement of 11.33 points (p < 0.001) on total CTRS score. Llama 3.1 8b had the strongest performance (mean CTRS score 67.86 +/- 7.24), followed by Qwen 2.5 7b (64.28 +/- 9.55) and Mistral 7b v0.3 (64.17 +/- 9.79), with these differences between models being statistically significant. The CBT-tuned models were competent in implementing core CBT techniques and providing empathetic responses, however, there were limitations observed in agenda adherence, exploration depth and long-context coherence. This study establishes that CBT specific fine-tuning can effectively encode therapeutic competencies in small LLMs, though significant technical and ethical considerations must be resolved prior to clinical deployment.

翻译：认知行为疗法（CBT）是治疗重度抑郁症的一种成熟且基于证据的方法。然而，个体在获取CBT方面存在显著障碍，包括费用高昂、治疗师稀缺以及社会污名化。本研究探讨了微调小型开源权重大型语言模型（LLMs）以提供抑郁症CBT的可行性。利用由Nous Research微调的Llama 3.1 405b生成的58组合成CBT对话记录，我们对三个模型进行了微调：Mistral 7b v0.3、Qwen 2.5 7b和Llama 3.1 8b。通过修改版的认知治疗评定量表（CTRS）评估了CBT保真度。所有微调模型均相互比较，并与它们的指令微调变体进行对比。为评估模型性能，生成了模拟患者对话记录，其中指令微调模型和CBT微调模型扮演治疗师角色，DeepSeek-V2.5扮演患者角色。这些模拟对话记录由Gemini 1.5 Pro-002使用修改版CTRS进行评估。我们的研究结果表明，CBT微调模型在CTRS总分上显著优于其指令微调对应模型，平均提升11.33分（p < 0.001）。Llama 3.1 8b表现最佳（平均CTRS得分67.86 +/- 7.24），其次是Qwen 2.5 7b（64.28 +/- 9.55）和Mistral 7b v0.3（64.17 +/- 9.79），模型间的差异具有统计学意义。CBT微调模型能够有效实施核心CBT技术并提供共情回应，但在议程遵循、探索深度和长上下文连贯性方面存在局限性。本研究证实，针对CBT的特定微调可以有效将治疗能力编码到小型LLMs中，但在临床部署前仍需解决重大的技术和伦理问题。