300 rows from the MS_MARCO Dataset reworked for training via Direct Preference Optimization. The prompt format is for the Mistral Instruct Models.
Original Dataset is not mine and licencing driven by licencing of original dataset. Posted this as it may be of use to others.