A team wants to submit a prompt dataset to Bedrock, run model inference, score the responses automatically, and view aggregated results in reports. Which Bedrock capability should it use?

Question

Accepted Answer

A. Automatic model evaluation. Automatic model evaluation in Bedrock runs inference on prompt datasets, scores the outputs with evaluation algorithms, and presents aggregated results. Prompt attack detection is a safety control, not an evaluation workflow. Automatic model evaluation is therefore the correct answer.

A team wants to submit a prompt dataset to Bedrock, run model inference, score the responses automatically, and view aggregated results in reports. Which Bedrock capability should it use?

Related Questions

Discussion