sebastian
Joined 18 years ago
Joined 18 years ago
The thing I'd be curious to separate out is ratio vs density. The fiction examples were positive, but a lot of the tokens are still spent on normal story work. The targeted examples put much more of the training signal on the AI being in the relevant situation and choosing against the bad option.
That makes me think the next thing to test is not just the positive/negative mix, but how much of the data is actually about the failure mode.
They tried something close to that. Positive AI fiction and also a "virtuous character" setup. Those didn't seem to do nearly as well as the targeted examples.
What mattered, at least in this setup, was more specific. The model sees the actual failure-mode scenario, the bad action is available, and the example shows the AI choosing against it.
So this reads less like "nicer AI stories" to me, and more like inoculation.
But closeness makes it extremely difficult to find actual backdoors left on purpose. Open processors would be a first step in the right direction.
Postbox unfortunately doesn't support Linux, besides that while I don't mind paying licences for great software, it doesn't look like Postbox is fully opensource [1].
Most of the open source email clients I have tested require you to run a local webserver and access the mail using a web browser with very limited features. All I want/need is a desktop app that can be customized to work similar to Gmail, pulls and deletes emails from remote SMTP/IMAP servers and allows me to create backups locally.
Thunderbird gives me that.