Back to evaluations
Public evaluation

purdue_boundary_check

Evaluates RAG retrieval precision, compliance violation detection, and query execution latency.

Evaluation type
task based
Challenge
LlamaIndex Purdue Model OT/IT RAG Pipeline with Together AI
Difficulty
Intermediate
Rigor
Unspecified

Evaluation overview

How the linked challenge is judged: tasks, benchmarks, and criteria count.

Tasks
1
Benchmarks
0
Criteria
0

Task templates

Inputs and expected outputs.

Task 1

purdue_boundary_check

Query the RAG system with proposed data pipeline paths to flag Purdue layer boundary violations.

Input format

JSON query specifying source_level, target_level, and proposed protocol

Output format

JSON with allowed (bool), boundary_violations (list), context_retrieved (string), search_time_ms (int)