-
Notifications
You must be signed in to change notification settings - Fork 3.3k
Closed
Description
If I understand correctly, the oasst1 dataset only contains ranking data that has been aggregated by "Tideman’s method" (according to the paper). As there are other aggregation schemes and little research on their different merits in conjunction with RLHF, are there plans to include more detailed ranking information (i.e. counts for each unique personal ranking) in future dataset releases?
Metadata
Metadata
Assignees
Labels
No labels