Add fixing DPO step 3 input
#4
by kuznetcova - opened
Output of prepare_for_selfcorrection.py on 5248 generated/ files, input for eval_math.py for self-correction generations (intermediate output of data_pipeline\fixing_dpo\step3_selfcorrect.sh basically)
Merged after resolving the .gitattributes LFS-rule conflict while preserving both uploaded files.
Alepach changed pull request status to merged
Alepach deleted the
refs/pr/4 ref