Comment by alienbaby 6 days ago can they be considered to have compressed the entirety of their training data into their weights? 4 comments alienbaby Reply nomel 6 days ago If you consider them lossy compression, then yes.The goal would be to find the minimum model that, with a fixed seed, would exactly reproduce your text. anothereng 6 days ago So I could hide information in a model basically nomel 6 days ago I think this would be the exact opposite. This model would be optimized to only output your information. 1 reply →
nomel 6 days ago If you consider them lossy compression, then yes.The goal would be to find the minimum model that, with a fixed seed, would exactly reproduce your text. anothereng 6 days ago So I could hide information in a model basically nomel 6 days ago I think this would be the exact opposite. This model would be optimized to only output your information. 1 reply →
anothereng 6 days ago So I could hide information in a model basically nomel 6 days ago I think this would be the exact opposite. This model would be optimized to only output your information. 1 reply →
nomel 6 days ago I think this would be the exact opposite. This model would be optimized to only output your information. 1 reply →
If you consider them lossy compression, then yes.
The goal would be to find the minimum model that, with a fixed seed, would exactly reproduce your text.
So I could hide information in a model basically
I think this would be the exact opposite. This model would be optimized to only output your information.
1 reply →