Analytics
We are adding some simple analytic features to the LLM grader to provide a real‑time window into all grading activity performed on the portal. The analytics are intended for instructors, TAs, and maintainers who need visibility into:
- grading results
- timeouts and errors
- latency and performance
- raw prompts and model behavior
- debugging and operational analytics
We are hoping that we can learn from this data to improve future LLM-based teaching systems.
Right now we just a have a single page: