3 ms·
We Found an Neuron in GPT-2
- cottenio 4y agoThis is one of the first articles I’ve read with a decent attack on reverse engineering the black boxes of neural networks. I particularly appreciate the use of corrupted prompts for isolating behaviors.
- danjc 4y agoCame here to say something similar. It seems to me that being able to determine how specific neurons are affecting outputs will be crucial to future optimization.