The Phi-4 research team called the human-generated dataset “organic.” Indeed, it must be 100% pesticide-free.

After finding pivotal token, they created synthetic preference dataset which includes (context + accepted token) with positives and (context + rejected token) as negatives.

According to their technical report, Phi-4 uses different approach than Marco-o1 has: careful data curation and enhanced DPO (post training).

arxiv.org/abs/2412.08905

Kaggle Startar Packなるものを紹介された
go.bsky.app/VFijtNt