They is no chance they are the one training it. It costs hundreds of millions to get a descent model. Seems like they will be using mistral, who have scrapped pretty much 100% of the web to use as training data.
Comment on Proton mail launches LLM and crypto wallet
ChilledPeppers@lemmy.world 3 months agoYeah, but their llm doesn’t even disclose where it gathered its training data, very sus.
L_Acacia@lemmy.one 3 months ago
Petter1@lemm.ee 3 months ago
It has to be very good with porn stuff, in that case 🤔
0laura@lemmy.world 3 months ago
yes, you can download SD1.5 models that will generate all kinds of degenerate images for you and deneutered LLMs that will write the most disgusting smut you’ve ever seen. all of it locally, free and 100% private.
Petter1@lemm.ee 3 months ago
Nice 🧐 got to do some research
L_Acacia@lemmy.one 3 months ago
Mistral modèles don’t have much filter don’t worry lmao
hendrik@palaver.p3x.de 3 months ago
Techradar says it's based on the Mistral 7B large language model. But they should definitely disclose that kind of information. It's important to know how a tool works and what kind of mistakes, biases etc are to be expected when using it for important communication.