Fireworks explains supervised fine-tuning steps for training an open model

AIFireworks says its session stage features a talk on training a small decision model with SFT and using it to route agentic tasks across open models. NVIDIA presents at 11:30 AM PT and Microsoft at 12:00 PM PT.

AIFireworks' Prof_OZ explains how to use supervised fine-tuning (SFT) to train a behavior into an open model's weights, so prompts no longer need to repeat it. The steps are to pick one task, pair inputs with ideal outputs, train on Fireworks, test on new inputs, and compare against the original model. A full guide is linked in the post.