# HorizonBench: Long-Horizon Personalization with Evolving Preferences

arXiv: https://arxiv.org/abs/2604.17283 · HTML: https://mlsanity.com/paper/2604.17283/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 27 references that are in the map.
- First posted on arXiv 19 Apr 2026.

## Builds on (most cited first, 23 of 23)

- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [Evaluating Very Long-Term Conversational Memory of LLM Agents](https://mlsanity.com/paper/2402.17753/) (2024)
- [LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory](https://mlsanity.com/paper/2410.10813/) (2024)
- [LaMP: When Large Language Models Meet Personalization](https://mlsanity.com/paper/2304.11406/) (2023)
- [Knowledge Conflicts for LLMs: A Survey](https://mlsanity.com/paper/2403.08319/) (2024)
- [Know Me, Respond to Me: Benchmarking LLMs for Dynamic User Profiling and Personalized Responses at Scale](https://mlsanity.com/paper/2504.14225/) (2025)
- [Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs](https://mlsanity.com/paper/2502.09597/) (2025)
- [PersonaMem-v2: Towards Personalized Intelligence via Learning Implicit User Personas and Agentic Memory](https://mlsanity.com/paper/2512.06688/) (2025)
- [MemBench: Towards More Comprehensive Evaluation on the Memory of LLM-based Agents](https://mlsanity.com/paper/2506.21605/) (2025)
- [Evaluating Large Language Models as Generative User Simulators for Conversational Recommendation](https://mlsanity.com/paper/2403.09738/) (2024)
- [PersonaLens: A Benchmark for Personalization Evaluation in Conversational AI Assistants](https://mlsanity.com/paper/2506.09902/) (2025)
- [ChatBench: From Static Benchmarks to Human-AI Evaluation](https://mlsanity.com/paper/2504.07114/) (2025)
- [AI PERSONA: Towards Life-long Personalization of LLMs](https://mlsanity.com/paper/2412.13103/) (2024)
- [PersonaBench: Evaluating AI Models on Understanding Personal Information through Accessing (Synthetic) Private User Data](https://mlsanity.com/paper/2502.20616/) (2025)
- [Reliable LLM-based User Simulator for Task-Oriented Dialogue Systems](https://mlsanity.com/paper/2402.13374/) (2024)
- [Toward Multi-Session Personalized Conversation: A Large-Scale Dataset and Hierarchical Tree Framework for Implicit Reasoning](https://mlsanity.com/paper/2503.07018/) (2025)
- [PrefPalette: Personalized Preference Modeling with Latent Attributes](https://mlsanity.com/paper/2507.13541/) (2025)
- [PersonaFeedback: A Large-scale Human-annotated Benchmark For Personalization](https://mlsanity.com/paper/2506.12915/) (2025)
- [A LLM-based Controllable, Scalable, Human-Involved User Simulator Framework for Conversational Recommender Systems](https://mlsanity.com/paper/2405.08035/) (2024)
- [Towards Natural Personalization: Evaluating Long-Horizon Preference Following in Personalized User-LLM Interactions](https://mlsanity.com/paper/2603.04191/) (2026)
- [BenchPreS: A Benchmark for Context-Aware Personalized Preference Selectivity of Persistent-Memory LLMs](https://mlsanity.com/paper/2603.16557/) (2026)
- [PrefDisco: Benchmarking Proactive Personalized Reasoning](https://mlsanity.com/paper/2510.00177/) (2025)
- [Beyond Fixed Psychological Personas: State Beats Trait, but Language Models are State-Blind](https://mlsanity.com/paper/2601.15395/) (2026)

## Built on by (most cited first, 17 of 17)

- [Beyond Similarity: Trustworthy Memory Search for Personal AI Agents](https://mlsanity.com/paper/2606.06054/) (2026)
- [EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments](https://mlsanity.com/paper/2606.13681/) (2026)
- [MINTEval: Evaluating Memory under Multi-Target Interference in Long-Horizon Agent Systems](https://mlsanity.com/paper/2605.18565/) (2026)
- [MemTrace: Probing What Final Accuracy Misses in Long-Term Memory](https://mlsanity.com/paper/2606.17328/) (2026)
- [Best Friends, Not Forever: Evaluating Long-Horizon Persona Collapse and Behavioral Drift in AI Companions](https://mlsanity.com/paper/2607.28818/) (2026)
- [When Users Don't Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents](https://mlsanity.com/paper/2609.03467/) (2026)
- [Synthetic Computers at Scale for Long-Horizon Productivity Simulation](https://mlsanity.com/paper/2604.28181/) (2026)
- [PHASE-Tree: Modeling Character-State Evolution in Long-Horizon Role-Playing Dialogue](https://mlsanity.com/paper/2608.06975/) (2026)
- [Hypotheses-Guided Self Distillation for Continual Personalization](https://mlsanity.com/paper/2609.00251/) (2026)
- [Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States](https://mlsanity.com/paper/2609.15972/) (2026)
- [Beyond Oracle Communication: Benchmarking Interactive Intent Alignment Under Miscommunication and Evolving User Intent](https://mlsanity.com/paper/2609.38604/) (2026)
- [Recall Isn't Enough: Bounding Commitments in Personalized Language Systems](https://mlsanity.com/paper/2605.16712/) (2026)
- [REDDIT: Forgetting-Resistant Correction of Timestamp Drift in ASR via Replay-Based Distribution Editing](https://mlsanity.com/paper/2607.05364/) (2026)
- [CAPTURE: Disentangling Preference Drift from Memory Poisoning in Personalized LLM Agents](https://mlsanity.com/paper/2609.02265/) (2026)
- [SCLATE: a Substrate for Continual-Learning Agent Training and Evaluation](https://mlsanity.com/paper/2609.32391/) (2026)
- [RealCompanion: Benchmarking Human Understanding from Reasoning over Longitudinal Real-World Conversations](https://mlsanity.com/paper/2610.01780/) (2026)
- [AgentMemGate: Addressing Speculation Contamination in Conversational Assistant Memory](https://mlsanity.com/paper/2610.07707/) (2026)

## Abstract (arXiv)

User preferences evolve across months of interaction, and tracking them requires inferring when a stated preference has been changed by a subsequent life event. We define this problem as long-horizon personalization and observe that progress on it is limited by data availability and measurement, with no existing resource providing both naturalistic long-horizon interactions and the ground-truth provenance needed to diagnose why models fail. We introduce a data generator that produces conversations from a structured mental state graph, yielding ground-truth provenance for every preference change across 6-month timelines, and from it construct HorizonBench, a benchmark of 4,245 items from 360 simulated users with 6-month conversation histories averaging ~4,300 turns and ~163K tokens. HorizonBench provides a testbed for long-context modeling, memory-augmented architectures, theory-of-mind reasoning, and user modeling. Across 25 frontier models, the best model reaches 52.8% and most score at or below the 20% chance baseline. When these models err on evolved preferences, over a third of the time they select the user's originally stated value without tracking the updated user state. This belief-update failure persists across context lengths and expression explicitness levels, identifying state-tracking capability as the primary bottleneck for long-horizon personalization.