The lab that ships an MMLU score rarely says how it was measured
GPT-4's headline 86.4% on MMLU shipped with its five-shot count and little else — not the prompt template, not the scoring rule, not the harness that turned 57 subjects of multiple choice into a single number.
An attention head is a weighted average that computes its own weights
Nine dot products across three tokens compute one attention head's output, and the same nine show why long context costs what it does and why the softmax weights are not, by themselves, an explanation.
Chinchilla beat a model four times its size on the same compute
The paper's tables buy a training-compute-optimal ratio near twenty tokens per parameter, evidenced three ways, but not the settled law the field now quotes or the inference cost that makes everyone train past it.
The people who can halt a dangerous model launch all work for the lab shipping it
Anthropic, OpenAI, and Google DeepMind each publish a rule for pausing a dangerous model, and in all three the final call belongs to the CEO or an officer the CEO appoints.