How to Answer and Grade Questions

Pre-requisites

Before using the LLM grader, you will need to register an OpenAI API key. Loading the OpenAI API key takes just a moment and lets you control your own usage and costs — nothing is stored on the server, and you can turn it off anytime.

Your instructor should provide you with the URL where they have deployed the class. Typically, this website is on render.com. For example, for the Introduction ot Hardware Design class at NYU, uses the following render web portal

Grade View

Once you have set the OpenAI API key and a course URL, go to the Grade View where you’ll spend most of your time. To open it, select File → Switch View → Grade. In this view, you can read the question, write your answer, and get instant feedback from the LLM grader. No mystery, no hidden steps — just a clean loop of try → grade → improve.

  • Select a unit from the Unit dropdown.
  • Select a question from the Questions dropdown.
  • In the Question panel (left side on desktop, or the Question tab on mobile), you will see the problem and any diagrams or code snippets.
  • Below the question, the Your Solution panel contains a composer where you enter your answer. You can write in plain English, math, or short code fragments — whatever the question calls for. The grader is flexible, but clearer answers usually get clearer feedback.

The Solution Composer

The solution entry area is a chat-style composer with three parts:

  • ＋ (attach) button — attach one or more image files (photos, sketches, screenshots) to your answer.
  • Text area — type your answer here. The area expands as you type, and if your cursor is in the box you can paste an image directly from the clipboard.
  • Grade button — submit your answer for grading.

You can also press Ctrl+Enter (or ⌘+Enter on Mac) to grade without reaching for the mouse.

Attaching Images

If your answer includes hand-written work, circuit diagrams, or plots, attach them as images:

  1. Click or tap the ＋ button inside the composer.
  2. Select one or more image files from your device.
  3. Thumbnails appear above the text area confirming the images are attached.
  4. To remove an image before grading, click the × on its thumbnail.

You can also paste an image from the clipboard while the text area is focused. This is useful for screenshots or copied whiteboard work.

Attached images are sent to the LLM along with your text, so the grader can see both. Up to three images may be attached per question.

Grading

Once you have typed your answer (and optionally attached images), click the Grade button in the composer. Within a few seconds (typically 5-10 seconds) you will see:

  • Summary — a quick, student-friendly explanation of how your answer compares.
  • Full Explanation — a more detailed breakdown of the reasoning.

This feedback is meant to help you understand why something is correct or incorrect, not just whether you got it right.

You can revise your answer as many times as you like. If the feedback points out something you missed, edit your answer and grade again. Many students use this loop to check their understanding before submitting the final version.

Saving and Loading Your Work

Your latest answer and grading results stay on the page, so you can generally switch views or come back later without losing anything. If you want to save your results to a file, select the Grade → Save Results… menu option. This creates a JSON file with:

  • Your solution text
  • Any attached images
  • The feedback from OpenAI
  • The full explanation
  • The grade result

The file is downloaded to your Downloads folder. You can store it anywhere and reload it later with Grade → Load Results….

If you’re ready to see how you’re doing across the whole assignment, head over to the Dashboard. When you’re satisfied with your answers, follow the instructions there for submitting to Gradescope.

This site uses Just the Docs, a documentation theme for Jekyll.