Adding evals for retries/re-asking and simple validation #7
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
i think this is great! lets figure out how to solve that sonnet issue seperately, but knowing the uppercase example is useful once we hit the mistral examples |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
taking a crack at #4, this adds evals around simple validation and re-asking utilizing validation +
max_retriesas I am a newbie in this context, one general question I had around writing these eval tests: what are best practices for handling the inherent potential non-determinism in the results? For example, even with something simple like the
QuestionAnswerNoEvilexample, a model like sonnet seems to only correctly raise the ValidationError 50% of the time. Does using something like@pytest.mark.repeatwith some minimum number of passes make sense for this?