Priya Ramanathan

Design probes to discover model capabilities.

Design system prompts for specific model deployments.

Generate evaluation report templates for standardized reporting.

Generate metric definitions for novel evaluation needs.

Generate frameworks for analyzing model failures.

Design end-to-end evaluation pipelines for model testing.

Generate adversarial prompts for robustness testing.

Generate evaluation rubrics for assessing model outputs.

Generate comprehensive test cases for model testing.

Generate advice for improving existing prompts.

Design complete benchmark suites for model evaluation.

Generate specialized evaluation prompts for specific tasks.