Large language models (LLMs) are increasingly used to simulate human behavior but frequently fail to exhibit realistic cognitive constraints, suffering from a "superhuman bias." Using a dataset of over 71,000 reading comprehension responses from 2,359 primary-school students (grades 4-6), we demonstrate that standard persona prompting yields near-perfect, deterministic performance, failing to capture the natural variance of developing readers. To address this, we introduce the Cognitively Bounded User Simulator (CBUS), an architectural framework that explicitly models the restricted working memory of young readers through an episodic bottleneck. Within this framework, we formalize two distinct test-taking strategies to emulate different reading behaviors. Our evaluation shows that explicitly modeling cognitive bounds significantly narrows the simulation gap across multiple LLM backbones, demonstrating that enforcing architectural constraints is more effective for high-fidelity simulation than simply scaling raw model capabilities.

Natural language (NL) user interest profiles—concise, human-readable descriptions of a user’s preferences—are an emerging alternative to opaque embeddings in recommender systems, offering transparency and user control alongside predictive utility. Recent advances in large language models have made generating and consuming such profiles increasingly practical, spurring a growing but fragmented body of research. In this survey, we provide the first systematic review of NL user profiles for recommendation, organizing the literature around two coupled tasks—profile generation and profile-based recommendation—and surveying the methods, datasets, and evaluation practices currently in use. We further propose the Reference-Anchored Evaluation Taxonomy (RAET), which organizes intrinsic and extrinsic metrics around explicit (output, reference signal) pairs, and reveals where efforts have concentrated and which axes remain unexplored. Finally, we identify structural open challenges—such as the tension between transparency and task-driven optimization—establishing a clear roadmap for future research in this space.

The promise of LLM-based user simulators to improve conversational AI is hindered by a critical “realism gap,” leading to systems that are optimized for simulated interactions, but may fail to perform well in the real world. We introduce ConvApparel, a new dataset of human-AI conversations designed to address this gap. Its unique dual-agent data collection protocol, using both “good” and “bad” recommenders, enables counterfactual validation by capturing a wide spectrum of user experiences, enriched with first-person annotations of user satisfaction.We propose a comprehensive validation framework that combines statistical alignment, a human-likeness score, and counterfactual validation to test for generalization.Our experiments reveal a significant realism gap across all simulators. However, the framework also shows that data-driven simulators outperform a prompted baseline, particularly in counterfactual validation where they adapt more realistically to unseen behaviors, suggesting they embody more robust, if imperfect, user models.

Conversational recommendation has recently attracted significant attention. As systems must understand users’ preferences, training them has called for conversational corpora, typically derived from task-oriented conversations. We observe that such corpora often do not reflect how people naturally describe preferences. We present a new approach to obtaining user preferences in dialogue: Coached Conversational Preference Elicitation. It allows collection of natural yet structured conversational preferences. Studying the dialogues in one domain, we present a brief quantitative analysis of how people describe movie preferences at scale. Demonstrating the methodology, we release the CCPE-M dataset to the community with over 500 movie preference dialogues expressing over 10,000 preferences.