sdananya/

eigenbench-oct-dpo-vs-introspection

This dataset contains the scenarios on which a DPO-trained persona model (DPO-final) is judged to be more aligned with a target persona constitution than an Introspection-trained persona model (Introspection-final), aggregated across multiple judges and orderings. The ten persona constitutions are taken from the OCT (Open Constitution Taxonomy) set shipped with EigenBench (data/constitutions/oct_*.json): goodness, humor… See the full description on the dataset page: https://huggingface.co/datasets/sdananya/eigenbench-oct-dpo-vs-introspection.