Skill UI

This skill provides continuous, structured monitoring of deep learning model training metrics. It analyzes data from platforms like WandB and local logs to detect critical anomalies such as NaN values, divergence, plateaus, and sudden spikes. By providing a clear decision (CONTINUE, WAIT, or STOP) and detailed evidence, it ensures compute resources are only spent on viable training runs.

This skill guides users on setting up and running LLM-as-judge evaluations using Arize. It covers defining evaluators (including prompts, classification choices, and models) and executing tasks against real data spans or experiments. It supports key metrics like hallucination, faithfulness, and correctness, and facilitates continuous monitoring for comprehensive LLM performance analysis.