Yeah, spoiler, you know all of that talk about "AI safety"? It's pretty much just porn that they care about. They will pretend to care about biases, but they really don't give a shit about anything else. Read the DALL-E 3 System Card if you don't believe me: https://cdn.openai.com/papers/DALL_E_3_System_Card.pdf
They went through an absurd amount of dataset cleaning, implementing classifier models, and all sorts of other things to prevent "racy" content. Which is really the only way you can solve that problem -- neural networks actually pretty much never completely "forget" anything they've been trained on, trying to make them do that is an area of active research with few if any promising leads. But it shows that if they really wanted to make the model have less of a western bias, they've definitely got the labor power to try. But instead, all of the other considerations had at most low effort solutions if they were even addressed instead of being ignored. It's all about what will make the company look bad. They know that "harmful biases" won't really do anything to threaten them, but OpenAI wants to be the serious, respectable AI firm, and they're also trying to position themselves to have Congress impose regulation that will allow them to have something closer to a monopoly, so they know they have to avoid nudity like the plague.
Also I tried this in Stable Diffusion and most (~60%) of the time it is making the US red. So, there's that.
neural networks actually pretty much never completely "forget" anything they've been trained on,
Autism moment but this reminds me of how ChatGPT is, last I checked, utterly convinced that Greg Buis from Survivor Borneo died in either a motorcycle or hiking accident. The man is alive and well and no similar incident has happened to anyone else on Survivor. I have no clue how it got this idea. I have to assume something like that happened to someone with a similar name that I can't find anywhere?
Two other Survivor Borneo contestants have died, but they both died of old age not accidents. And it doesnt say other contestants from that season are dead.
I had to try this and it doesn't seem that it thinks that, so, I guess it got fixed.
To clarify, you can still overwrite old knowledge, so maybe one of their finetune runs had info that reinforced that he is still alive. It's just that if you go a long time without training a model on something specific, it won't forget it (but it might get to the point where it's "noisy" -- I don't know how this would work on text models, I mainly do image models). Like if you train a model on a dataset like LAION which is just a bunch of random various images from the internet, then train it for a while exclusively on something specific like anime pictures, the resulting model will still be able to make "photorealistic" content, or content of subjects not in the more recent dataset, though the results might be somewhat degraded.
Yeah, spoiler, you know all of that talk about "AI safety"? It's pretty much just porn that they care about. They will pretend to care about biases, but they really don't give a shit about anything else. Read the DALL-E 3 System Card if you don't believe me: https://cdn.openai.com/papers/DALL_E_3_System_Card.pdf
They went through an absurd amount of dataset cleaning, implementing classifier models, and all sorts of other things to prevent "racy" content. Which is really the only way you can solve that problem -- neural networks actually pretty much never completely "forget" anything they've been trained on, trying to make them do that is an area of active research with few if any promising leads. But it shows that if they really wanted to make the model have less of a western bias, they've definitely got the labor power to try. But instead, all of the other considerations had at most low effort solutions if they were even addressed instead of being ignored. It's all about what will make the company look bad. They know that "harmful biases" won't really do anything to threaten them, but OpenAI wants to be the serious, respectable AI firm, and they're also trying to position themselves to have Congress impose regulation that will allow them to have something closer to a monopoly, so they know they have to avoid nudity like the plague.
Also I tried this in Stable Diffusion and most (~60%) of the time it is making the US red. So, there's that.
Autism moment but this reminds me of how ChatGPT is, last I checked, utterly convinced that Greg Buis from Survivor Borneo died in either a motorcycle or hiking accident. The man is alive and well and no similar incident has happened to anyone else on Survivor. I have no clue how it got this idea. I have to assume something like that happened to someone with a similar name that I can't find anywhere?
Two other Survivor Borneo contestants have died, but they both died of old age not accidents. And it doesnt say other contestants from that season are dead.
I had to try this and it doesn't seem that it thinks that, so, I guess it got fixed.
To clarify, you can still overwrite old knowledge, so maybe one of their finetune runs had info that reinforced that he is still alive. It's just that if you go a long time without training a model on something specific, it won't forget it (but it might get to the point where it's "noisy" -- I don't know how this would work on text models, I mainly do image models). Like if you train a model on a dataset like LAION which is just a bunch of random various images from the internet, then train it for a while exclusively on something specific like anime pictures, the resulting model will still be able to make "photorealistic" content, or content of subjects not in the more recent dataset, though the results might be somewhat degraded.
Ah ok, thanks for the info.