Open LuciusApollo opened 8 months ago
We would like to be able to make RIB graphs that exclude the first (BOS) token of the prompt from all basis and edge calculations. Could maybe implement optional masking of the activations over tokens to accomplish this.
Context:
We would like to be able to make RIB graphs that exclude the first (BOS) token of the prompt from all basis and edge calculations. Could maybe implement optional masking of the activations over tokens to accomplish this.