Inspect generated responses and media for prompt adherence, scenario quality and visible errors.

- Compare paired examples for requested details, coherent behavior and overall clarity.

- Explain the observed mismatch or strength that supports your choice.

Take part in English conversations with AI models when assigned to voice tasks.

- Record your ratings using project guidelines and flag examples that are difficult to judge.