Open-PerfectBlend is an open-source reproduction of the instruction dataset introduced in the paper "The Perfect Blend: Redefining RLHF with Mixture of Judges".
It's a solid general-purpose instruction dataset with chat, math, code, and instruction-following data.
openbmb/UltraInteract_sft
288,579β¦ See the full description on the dataset page:
https://huggingface.co/datasets/mlabonne/open-perfectblend.