◈ gemma WebGPU ● loading
LAYERS
embed
lm_head
Select a layer to inspect its attention weights
GQA = grouped-query attention · SWA = sliding window attention
Context