Beyond One Output: Visualizing and Comparing Distributions of Language Model Generations
For researchers and users of language models, GROVE addresses the problem of over-generalizing from single outputs by providing a way to visualize and compare distributions of generations, though the gains are incremental over existing methods.
The paper introduces GROVE, an interactive visualization that represents multiple language model generations as overlapping paths through a text graph to reveal distributional structure. In three crowdsourced studies (N=47, 44, 40), graph summaries improved structural judgments like assessing diversity, while direct output inspection remained better for detail-oriented questions.
Users typically interact with and evaluate language models via single outputs, but each output is just one sample from a broad distribution of possible completions. This interaction hides distributional structure such as modes, uncommon edge cases, and sensitivity to small prompt changes, leading users to over-generalize from anecdotes when iterating on prompts for open-ended tasks. Informed by a formative study with researchers who use LMs (n=13) examining when stochasticity matters in practice, how they reason about distributions over language, and where current workflows break down, we introduce GROVE. GROVE is an interactive visualization that represents multiple LM generations as overlapping paths through a text graph, revealing shared structure, branching points, and clusters while preserving access to raw outputs. We evaluate across three crowdsourced user studies (N=47, 44, and 40 participants) targeting complementary distributional tasks. Our results support a hybrid workflow: graph summaries improve structural judgments such as assessing diversity, while direct output inspection remains stronger for detail-oriented questions.