Improving the review feature

Open
#33 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
35/100
Issue type
Feature
Clarity
Needs clarification
Activity status
Quiet
Tech stack
jupyter, typescript
Domain
ai

Research direction

Start by reading PR #28 and tracing the review feature’s current comparison path. Reproduce cases where the tutor misses differences, then evaluate the deterministic diff approach alongside model temperature and system-prompt effects. Done means differences between initial and student-written code are detected consistently for the affected models.

Written by the indexing model from the issue text.

Description

One of the main issues with the review feature right now is its inconsistency as mentioned in PR #28 (at least for some models): sometimes the tutor can't spot the difference between the initial code and the student written code. This could be due to the model not being good enough, the temperature and system prompt could also have an effect.
One solution I tried implementing for this is a deterministic diff calculator so that we don't have to rely on the LLM to find the differences. Using our own comparison tool could be a source of error, but I think this feature still needs to be improved some way in this aspect.

Dominant language
TypeScript
Stars
4
Forks
4
PR merge metrics
No merged PRs in 30d

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from QuantStack/jupyter-ai-tutor

All issues in QuantStack/jupyter-ai-tutor

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.