Use Japanese prompt instead of English for JNLI #110
Merged
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
mrorii
requested review from
mkshing and
polm-stability
and removed request for
jon-tow
polm-stability
approved these changes
Oct 25, 2023
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Outdated suggestions cannot be applied.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.
Overview
This PR replaces the English words
entailment,contradiction, andneutralin the prompt for the JNLI task with the Japanese words含意,矛盾, and中立. The purpose behind this is to guarantee that we evaluate the Japanese capability of the language models (and not necessarily English capabilities).Note that the task version has been bumped from 1.1 to 1.2, since this modifies the task definition.
Details
The following table shows the balanced accuracy for JNLI across several models.
As shown, prior to this PR, many models have a score of 33.33, which is essentially the same as random chance (given that there are 3 classes to choose from). After this PR, none of the models have a score of 33.33.