microsoft / microsoft/promptpex

Feature request: Use the LLM to tag rules that are checkable using code

Open
#9 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
TeX
Stars
164
Forks
21
Avg merge
1d 13h
Merged PRs (30d)
1

Description

We can leverage the LLM to determine what rules are checkable with code and then generate code to apply them to the test cases. Since there are probably a limited number of possible code checkers, we could have a set of pre-defined checkers that the LLM would just have to choose between (possibly).

We talked about this feature but didn't have time to add it.

Checkers can be generated both as input filters (rejecting inputs that don't meet the requirements) and as output filters.

Contributor guide

No contributing guide indexed for this repository

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.

Research direction

No files, tests, or entry points are identified. Start by mapping the existing rule and test-case evaluation flow, then determine how predefined checkers could be selected or generated for both input and output filtering; done means those checkers are applied to test cases as described.

Written by the indexing model from the issue text.

Assessment

Domain
ai, testing
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.