kishorepurushothaman07 / exampilot-study-plan-evaleval
One-sentence benchmark measuring whether an AI agent creates feasible, syllabus-grounded exam study plans, respects time limits, handles missing information, recovers from missed sessions, and avoids unsupported score guarantees.
educationstudy-planningAI-agents
1
README.md553 B
ExamPilot Study-Plan Eval
An evaluation set for testing an AI agent's syllabus-to-study-plan behaviour.
- Cases: 8
- Suggested pass bar: 80% (at least 7/8 cases)
- Dataset:
dataset.jsonl(one JSON object per line) - Rubric:
rubrics/evaluation-rubric.md - Licence: MIT
This package defines test cases only. No model was run while preparing it, so it contains no claimed evaluation results. Run all cases with the published ExamPilot skill installed, record the agent/model and actual pass count, and retain short evidence notes for each result.