ThoughtTrace: Understanding User Thoughts in Real-World LLM Interactions

Conversational AI has now reached billions of users, yet existing datasets capture only what people say, not what they think. We introduce $\textit{ThoughtTrace}$, the first large-scale dataset that pairs real-world multi-turn human–AI conversations with users' self-reported thoughts: their reasons for sending prompts and reactions to assistant responses. $\textit{ThoughtTrace}$ comprises 1,058 users, 2,206 conversations, and 10,556 thought annotations collected across 20 language models. Our analysis shows that $\textit{ThoughtTrace}$ captures long-horizon, topically diverse interactions, and that thoughts are semantically distinct from messages, difficult for frontier LLMs to infer from context, and tied to the conversation stages. We further demonstrate the utility of thoughts for downstream modeling. First, thoughts improve user-behavior prediction as conditions for inference. Second, thought-guided rewrites provide fine-grained alignment signals for training personalized assistants. Together, $\textit{ThoughtTrace}$ establishes user thoughts as a new data modality for studying the thought dynamics behind human--AI interaction and provides a foundation for building assistants that better understand and adapt to users' latent goals, preferences, and needs.