Language models, fine-tuning, and small models.
Ekcron explores how language models work, how fine-tuning and distillation shape them, and how small models can match far larger ones — through writing and experiments.
Understanding LLMs
How large language models actually behave — their capabilities, limits, and the mechanics underneath.
Fine-tuning
Adapting open models to specific tasks: the methods, the trade-offs, and what holds up in practice.
Small models
Getting big-model quality from compact models through distillation — and how to tell whether it worked.
Research
All research →Latest writing
All posts →What fine-tuning actually changes
Fine-tuning doesn't teach a model new facts so much as it reshapes behaviour. Here's a practical way to think about what moves and what doesn't.
Why small models are getting good
Compact models keep closing the gap with the giants on narrow tasks. A short look at what's driving it and where the limits still are.