RAG Evaluation

Load example:

User Input *

Response

Reference (Ground Truth)

Retrieved Contexts (JSON array of strings)

Metric

Judge Model

Load example:

User Input *

Retrieved Contexts * (JSON array of strings)

Reference (Ground Truth)

Reference Contexts (JSON array of strings)

Metric

Judge Model

Top K

Expected JSON format: Each case should contain case_id (string), user_input (string), and optionally response, reference, retrieved_contexts. Max 15 cases.

Drag & drop a JSON file here

or click to select a file

Metric

Judge Model

Expected JSON format: Each case should contain case_id (string), user_input (string), retrieved_contexts (array of strings, required), and optionally reference, reference_contexts. Max 15 cases.

Drag & drop a JSON file here

or click to select a file

Metric

Judge Model

Top K