No the poorly redacted ones were real, from the first release. They were not redacted by the DOJ, they were old court documents. Afaik all of them were also already released unredacted.

It’s only an issue with LLMs. And it’s because they’re generative, text completion engines. That is the actual learned task, and it’s a fixed task.

It’s not actually a chat bot. It’s completing a chat log. This can make it do a whole bunch of tasks, but there’s no separation of task description and input.