Priming, Path-dependence, and Plasticity: Understanding the molding of user-LLM interaction and its implications from (many) chat logs in the wild

2026-05-07

AI Summary

This study addresses a critical gap in understanding how users interact with large language models (LLMs) in real-world settings, particularly how prior experience and early exploration shape long-term behavioral patterns. Analyzing 140,000 longitudinal chat logs from 7,955 global users through behavioral trajectory modeling, textual pattern recognition, and correlational analysis, the work uncovers an “agency paradox”: despite the open-ended input space of LLMs, user exploration remains highly constrained. The study provides the first empirical evidence of strong path dependence, demonstrating that interaction patterns rapidly solidify during initial encounters and that early exploratory behavior significantly predicts long-term retention and linguistic repetitiveness. Furthermore, it identifies a dual dynamic mechanism driving behavioral evolution—task type and model updates—operating in parallel to shape sustained user engagement.

Abstract

User interactions with LLMs are shaped by prior experiences and individual exploration, but in-lab studies do not provide system designers with visibility into these in-the-wild factors. This work explores a new approach to studying real-world user-LLM interactions through large-scale chat logs from the wild. Through analysis of 140K chatbot sessions from 7,955 anonymized global users over time, we demonstrate key patterns in user expressions despite varied tasks: (1) LLM users are not tabula rasa, nor are they constantly adapting; rather, interaction patterns form and stabilize rapidly through individual early trajectories; (2) Longitudinal outcomes, such as recurring text patterns and retention rates, are strongly correlated with early exploration; (3) Parallel dynamics are present, including organizing expressions by task types such as emotional support, or in response to model-version updates. These results present an ``agency paradox'': despite LLM input spaces being unconstrained and user-driven, we in fact see less user exploration. We call for design consideration surrounding the molding procedure and its incorporation in future research.

Problem

• priming
• path-dependence
• plasticity
• user-LLM interaction
• chat logs

Innovation

• priming
• path-dependence
• plasticity
• user-LLM interaction
• chat logs