3 ms·
Aren't they basically safe as long as future AI models are still basically multiplying tensors of floats?
by ClassyJacket 2y ago
Aren't they basically safe as long as future AI models are still basically multiplying tensors of floats?
- cma 2y agoTraining uses other operations for normalizing, etc.