So, I'm using a vision LLM to do content evals on the output of a text-to-image model. The output is a series of questions with boolean values indicating if the image violated that content guardrail or not. see image
My question is, how can i use Arize dashboard to visualize this data? Specifically I want to display the prediction_id and other metadata (city, country, prompt, etc) for rows that show a content violation. I'm having trouble setting that up in the dashboards
Alternately, should I be structuring my data a different way to align with the built in dashboard use cases like time series? (maybe setting up a Monitor is a better way to go)?
