The AI Evaluation Record

1. What is the purpose of the `ops.AiEvaluations` table in LyraLearn?

To store raw logs and traces To capture a durable record of each AI call To log errors from AI model calls To analyze user behaviors during AI interactions

2. Why is the `PromptHash` column important in the `ops.AiEvaluations` table?

It serves as a unique identifier for each AI model It allows grouping of identical prompts to track changes It summarizes the cost of each AI call It records the confidence level of responses

3. What does a boolean flag in the `Refused` column signify?

The AI call completed successfully The AI declined to answer the request The model returned an error The response was generated with high confidence

4. How does the `LatencyMs` column contribute to analysis in the `ops.AiEvaluations` table?

It indicates the number of prompts processed It measures the duration of the AI call It tracks the total number of AI calls made It records the model's performance over time

Continue: Metrics That Matter → Review this lesson Retake quiz