6 docs tagged with "alignment"

Every type has an alignment: an address it must start on, always a power of two. char can live

Width and alignment are what turn a stream of values into a readable table, and the defaults differ

A pretrained model knows an enormous amount about language and often about the world — and none of it about how to behave. It has no notion that a question deserves a direct answer rather than a continuation of similar-looking text scraped from a forum. Fine-tuning is where capability turns into behaviour, and most complaints of "the model can't do X" are actually behaviour problems in disguise.

Data arranged at addresses that are multiples of its size. Required for correctness on some architectures, critical for performance on all.

This page is the low-level / ABI view: how padding affects a struct's on-the-wire byte layout,

Everything in this section so far assumed a reward function already exists. For "write a helpful, harmless response" — the actual goal behind training a modern chat model — no programmable reward function exists at all. RLHF is the answer: learn a reward function from human comparisons, then optimise against it. This is also, for most readers, where reinforcement learning actually shows up in practice.