3 ms·
I am glad for this because it shows what the whole fuss about "dangerous" was about. This is like an illusion that looks deceptive from afar, but is however tra
by return1 7y ago
I am glad for this because it shows what the whole fuss about "dangerous" was about. This is like an illusion that looks deceptive from afar, but is however transparent on any closer inspection. As long as it is impossible for deep learning to make sense, these things are not a danger.
- kristian_io 7y agoThis is just the 345M model, not the 1.5B one...
- return1 7y agothe samples from the full model did not make sense either, although they made less obvious grammar/syntax errors.