ML//model//GPT//GPT-2

- 1.5B parameters. "Too dangerous to release" — OpenAI staged the release in waves.

1.5B parameters. "Too dangerous to release" — OpenAI staged the release in waves.

No fine-tuning: pure pre-training, and it could write coherent paragraphs.

In retrospect, quaint. But at the time, the idea that a model could generate believable text was genuinely alarming to some people.