Fig.1

Concept

SFT

SFT stands for Supervised Fine-Tuning: you take a pretrained language model and continue training it on labeled input/output pairs, using the same next-token cross-entropy loss you already know from pretraining. The difference is the data.

The rest of “SFT” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.

Log in to unlock

← Back to the library