In ablation study, you show multi round revision works better than single round revision. You also show wizardcoder outperforms Pretrained model starcoder. However, it is unknown how much wizardcoder benefits from code alpaca or multi round prompt revision.
In ablation study, you show multi round revision works better than single round revision. You also show wizardcoder outperforms Pretrained model starcoder. However, it is unknown how much wizardcoder benefits from code alpaca or multi round prompt revision.