What fine-tuning can and can't fix

Why this matters

There is a clean dividing line that most people learn the hard way: fine-tuning is excellent at shaping behavior — style, tone, output format, following a niche instruction pattern, being concise — and bad at adding knowledge the base model never saw. If you fine-tune a model on 500 facts about your product, it will learn to sound like it knows your product while confidently making facts up, because supervised fine-tuning teaches the shape of answers, not a reliable fact store. Internalizing 'style and format: yes; new facts: no' saves you from the classic disappointment of a fine-tune that talks like an expert and hallucinates like an amateur.

Demo

The demo contrasts two training examples — one teaching a format/style the model can generalize, one trying to teach a specific fact — and explains why the first sticks and the second leaks into hallucination.

Try it yourself

1. List three things in your task that are behavior (style/format) and three that are knowledge (facts) — only the first list is a fine-tuning target.
2. Reason about why a fact taught via SFT goes stale, while the same fact in a RAG document can be updated instantly.
3. Find a borderline case (e.g. a fixed taxonomy of categories) and argue whether it's behavior or knowledge.
4. Predict what happens if you fine-tune on 50 product facts and then ask about a 51st product it never saw.

References

• Hugging Face — TRL (SFT) documentation
• OpenAI — Fine-tuning use cases