# Feel the Difference? A Comparative Analysis of Emotional Arcs in Real and LLM-Generated CBT Sessions

arXiv: https://arxiv.org/abs/2508.20764 · HTML: https://mlsanity.com/paper/2508.20764/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 33 references that are in the map.
- First posted on arXiv 28 Aug 2025.

## Builds on (most cited first, 11 of 11)

- [PsyQA: A Chinese Dataset for Generating Long Counseling Text for Mental Health Support](https://mlsanity.com/paper/2106.01702/) (2021)
- [Cognitive Reframing of Negative Thoughts through Human-Language Model Interaction](https://mlsanity.com/paper/2305.02466/) (2023)
- [Cactus: Towards Psychological Counseling Conversations using Cognitive Behavioral Theory](https://mlsanity.com/paper/2407.03103/) (2024)
- [HealMe: Harnessing Cognitive Reframing in Large Language Models for Psychotherapy](https://mlsanity.com/paper/2403.05574/) (2024)
- [REALTALK: A 21-Day Real-World Dataset for Long-Term Conversation](https://mlsanity.com/paper/2502.13270/) (2025)
- [Training Models to Generate, Recognize, and Reframe Unhelpful Thoughts](https://mlsanity.com/paper/2307.02768/) (2023)
- [NRC VAD Lexicon v2: Norms for Valence, Arousal, and Dominance for over 55k English Terms](https://mlsanity.com/paper/2503.23547/) (2025)
- [Emotion Dynamics in Movie Dialogues](https://mlsanity.com/paper/2103.01345/) (2021)
- [Evaluating Emotion Arcs Across Languages: Bridging the Global Divide in Sentiment Analysis](https://mlsanity.com/paper/2306.02213/) (2023)
- [The Emotion Dynamics of Literary Novels](https://mlsanity.com/paper/2403.02474/) (2024)
- [Generative Exaggeration in LLM Social Agents: Consistency, Bias, and Toxicity](https://mlsanity.com/paper/2507.00657/) (2025)

## Built on by (most cited first, 11 of 11)

- [Responsible Evaluation of AI for Mental Health](https://mlsanity.com/paper/2602.00065/) (2026)
- [DimABSA: Building Multilingual and Multidomain Datasets for Dimensional Aspect-Based Sentiment Analysis](https://mlsanity.com/paper/2601.23022/) (2026)
- [Auditing Stealth Sycophancy in Mental-Health Dialogue: Structured Clinical-State Diagnostics and Clean Matched Benchmarks](https://mlsanity.com/paper/2605.03472/) (2026)
- [Psych=eChat: An Empathic Framework Focused on Emotion Shift Tracking and Safety Risk Analysis in Psychological Counseling](https://mlsanity.com/paper/2601.12392/) (2026)
- [CBT-Audio: Evaluating Audio Language Models for Patient-Side Distress Intensity Estimation in CBT Session Recordings](https://mlsanity.com/paper/2605.17370/) (2026)
- [EmoTrack: Clinical-Semantic Modeling for Text-Based Depression Severity Estimation](https://mlsanity.com/paper/2605.22286/) (2026)
- [When Clients Stop Following: A Cognitive Conceptualization Diagram-driven Framework for Strategic Counseling](https://mlsanity.com/paper/2606.04389/) (2026)
- [From Triggers to Emotions: A CPM-Grounded Appraisal Multi-Agent for Dynamic Emotional Evolution in Persona-Based Dialogue](https://mlsanity.com/paper/2607.07824/) (2026)
- [LongCounsel-8: A Benchmark Suite for Longitudinal Depression Tracking from Multi-Session Counseling Dialogues](https://mlsanity.com/paper/2609.03507/) (2026)
- [CliniCIRCA: A Modular LLM Framework for Constructing Longitudinal Mental Health Patient Journeys from Raw EHR Narratives](https://mlsanity.com/paper/2609.19585/) (2026)
- [From Pattern Recognizers to Personalized Companions: A Survey of Large Language Models in Mental Health](https://mlsanity.com/paper/2609.25186/) (2026)

## Abstract (arXiv)

Synthetic therapy dialogues generated by large language models (LLMs) are increasingly used in mental health NLP to simulate counseling scenarios, train models, and supplement limited real-world data. However, it remains unclear whether these synthetic conversations capture the nuanced emotional dynamics of real therapy. In this work, we introduce RealCBT, a dataset of authentic cognitive behavioral therapy (CBT) dialogues, and conduct the first comparative analysis of emotional arcs between real and LLM-generated CBT sessions. We adapt the Utterance Emotion Dynamics framework to analyze fine-grained affective trajectories across valence, arousal, and dominance dimensions. Our analysis spans both full dialogues and individual speaker roles (counselor and client), using real sessions from the RealCBT dataset and synthetic dialogues from the CACTUS dataset. We find that while synthetic dialogues are fluent and structurally coherent, they diverge from real conversations in key emotional properties: real sessions exhibit greater emotional variability, more emotion-laden language, and more authentic patterns of reactivity and regulation. Moreover, emotional arc similarity remains low across all pairings, with especially weak alignment between real and synthetic speakers. These findings underscore the limitations of current LLM-generated therapy data and highlight the importance of emotional fidelity in mental health applications. To support future research, our dataset RealCBT is released at https://gitlab.com/xiaoyi.wang/realcbt-dataset.