In Azure Al Studio, you use Completions playground with the GPT-35 Turbo model.

You’re using GPT-3.5-Turbo in Completions playground .

You want the model to explain the given code.

The best approach is to directly add an instruction in the prompt (e.g., // what does function F do?).

This minimizes cost because:

You don’t need to upgrade to GPT-4 (more expensive).

You don’t need special formatting like function F(explanation).

Changing temperature only affects randomness, not clarity of explanation.

Microsoft Reference:

Prompt engineering with GPT models