Distillation

Search algorithm for generation that keeps top-k partial sequences; can improve likelihood but reduce diversity.

Stochastic generation strategies that trade determinism for diversity; key knobs include temperature and nucleus samp...

Scales logits before sampling; higher increases randomness/diversity, lower increases determinism.

Samples from the k highest-probability tokens to limit unlikely outputs.

Techniques that fine-tune small additional components rather than all weights to reduce compute and storage.

PEFT method injecting trainable low-rank matrices into layers, enabling efficient fine-tuning.

Reducing numeric precision of weights/activations to speed inference and reduce memory with acceptable accuracy loss.

Removing weights or neurons to shrink models and improve efficiency; can be structured or unstructured.