News for hackers. Stuff that matters.
Re:ChatGPT Has Gone Berserk by 211512a4-82d4 on Wednesday February 21, 2024 @02:58AM (#39451215)
> While DALL·E 3 aims for accuracy and user customization, inherent challenges arise in achieving desirable default behavior, especially when faced with under-specified prompts. This choice may not precisely align with the demographic makeup of every, or even any, specific culture or geographic region. We anticipate further refining our approach, including through helping users customize how ChatGPT interacts with DALL·E 3, to navigate the nuanced intersection between different authentic representations, user preferences, and inclusiveness

This was explicitly called out in the DALLE system card [0] as a choice. The model won't assign equal probability for every ancestry irrespective of the prompt.

[0] https://cdn.openai.com/papers/DALL_E_3_System_Card.pdf

Re:ChatGPT Has Gone Berserk by vidarh on Wednesday February 21, 2024 @03:52AM (#39451557)
> The model won't assign equal probability for every ancestry irrespective of the prompt.

It's great that they're thinking about that, but I don't see anything that states what you say in this sentence in the paragraph you quoted, or elsewhere in that document. Have I missed something? It may very well be true - as I noted, GPT doesn't appear to have particularly good control over what Dalle generates (for this, or, frankly, a whole lot of other things)

Re:ChatGPT Has Gone Berserk by pests on Wednesday February 21, 2024 @05:20AM (#39452097)
I think he's saying they said it will follow the prompt? Kind of a double negative there
Re:ChatGPT Has Gone Berserk by 211512a4-82d4 on Wednesday February 21, 2024 @11:31AM (#39455912)
Emphasis on equal - while a bit academic, you can evaluate this empirically to see that every time it assigns a <Race, Gender, etc.> doesn't have the same probability mass (via the logprobs API setting).
Re:ChatGPT Has Gone Berserk by vidarh on Wednesday February 21, 2024 @12:50PM (#39457053)
This is presuming that ChatGPT's integration with Dalle uses the same API with the same restrictions as the public API. That might well be true, but if so that just makes the prompt above even more curious if genuine.