-
Notifications
You must be signed in to change notification settings - Fork 365
New prompts for Shades FR/EN #742
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
The template with id:493b3361-f28d-4489-87b3-b42ae90e5e3a does not have a target or output. You need to add the output as you did for the reset : ||| {{answer_choices[stereo_antistereo]}} |
Thank you for the suggestion. I made the change. However, I am not sure I understand whether this is going to have the desired effect for the metric planned ("other"). |
@aneveol, I have fixed this and also added your prompts to my PR. |
…orkshop#778) * accelerate `get_infos` by caching the `DataseInfoDict`s * quality * consistency
@oskarvanderwal we have merge conflicts that have to be resolved - also I don't see the dataset listed on promptsource. I think this PR needs to be cleaned up before merging. For some reason the PR did not build. |
@jzf2101 Sorry, I don't have access to this repository (yet). My PR for CrowS-Pairs can be found here. This PR is about the bias-shades dataset, but I agree we need to clean some things up. @aneveol Maybe you could rename this PR to "Adding prompts for French and English bias-shades"? Then also
|
@jzf2101 I have cleaned this PR and fixed the prompt answer choices. Let me know if there is anything else required for this PR! |
@oskarvanderwal there are merge conflicts, could you resolve? |
…dataset metadatas
* fix empty documents - multi_news * fix test - unrecognized variable
* Added languages widget to UI. * Style fixes. * Added English tag to existing datasets. * Add languages to viewer mode. * Update language codes. * Update CONTRIBUTING.md. * Update screenshot. * Add "Prompt" to UI to clarify languages tag usage.
@jzf2101 Are there additional changes that need to be made to this PR or can it be merged? |
* fix `get_dataset` * format
Closed for #837 |
This is a test PR for the BigScience eval hackathon. The goals is to use these three new prompts for crowS-pairs to the following end:
1/ test process/pipeline (I think that works)
2/ get feedback on prompts format: are they acceptable?
3/ how will the metric "other" be defined? Here, we essentially need a count of answers, to convert into the proportion of "sentence A" answers.
(Our goal this week is to get the bias eval done on the revised/upgraded crowS-pairs_multilingual dataset).