Search papers, labs, and topics across Lattice.
To resolve the persistent conflict between immediate user intents and enduring traits in training-free LLM personalization, the authors formulate online adaptation as a latent preference tracing problem named HyperTrace. The framework maintains dual natural-language hypotheses for short-term intent and long-term preferences, dynamically updating their posterior weights across interaction turns via a Sequential Monte Carlo (SMC)-style reweighting process powered by an LLM surrogate choice model. Evaluated on the PRISM and PersonaMem-v2 benchmarks, HyperTrace consistently outperforms strong retrieval- and history-based baselines across response alignment, preference prediction, and multi-session profile consistency without modifying model parameters.
Instead of stuffing raw context with dialogue history or brittle retrieval vectors, treating user personalization as particle-filtered hypothesis tracing over natural language resolves the tension between fleeting intent and long-term preferences.
Personalized language models aim to adapt responses to individual users, whose preferences are often latent and revealed gradually through interaction. Existing training-free methods rely on stored histories or retrieved memories, but they often struggle to reconcile long- term preferences with short-term topic-specific needs. To address this issue, we propose HyperTrace, a training-free framework that formulates online personalization as latent preference tracing. HyperTrace maintains interpretable natural-language hypotheses over short-term intent and long-term preferences, and updates them through an SMC-style reweight process using an LLM-based surrogate choice model. By updating these hypotheses across turns and sessions, HyperTrace enables personalization without parameter updates. Experiments on PRISM and PersonaMem-v2 show that HyperTrace improves response alignment, preference prediction, and profile consistency over strong online baselines, demonstrating the effectiveness of tracing latent user preferences for robust personalization. Code and scripts are available in the repository: https://github.com/jiseshen/HyperTrace.
