A Deep User Interface for Exploring LLaMa
Divya Perumal, Swaroop Panda
TL;DR
This work tackles the interpretability and usability barrier of large language models by introducing a visual analytics-driven deep user interface that enables interactive exploration of key hyperparameters, specifically Top-P and frequency/presence penalties. The tool, built around the LLaMA-7B API, uses an intuitive UI with a Top-P knob, a 2D penalty-scatter, prompts, and history to compare outputs across parameter settings, and is evaluated through a user study with 10 participants, yielding favorable feedback on aesthetics, layout, and navigability. Quantitative results show high usability ratings, while qualitative feedback highlights the value of visual exploration in understanding model behavior and the importance of accessible hyperparameter descriptions. The study provides preliminary actionable insights and points to future work for broader hyperparameter coverage and more extensive validation, underscoring the potential of VA-driven interfaces to make HL models more transparent and user-friendly in practice.
Abstract
The growing popularity and widespread adoption of large language models (LLMs) necessitates the development of tools that enhance the effectiveness of user interactions with these models. Understanding the structures and functions of these models poses a significant challenge for users. Visual analytics-driven tools enables users to explore and compare, facilitating better decision-making. This paper presents a visual analytics-driven tool equipped with interactive controls for key hyperparameters, including top-p, frequency and presence penalty, enabling users to explore, examine and compare the outputs of LLMs. In a user study, we assessed the tool's effectiveness, which received favorable feedback for its visual design, with particular commendation for the interface layout and ease of navigation. Additionally, the feedback provided valuable insights for enhancing the effectiveness of Human-LLM interaction tools.