Model: GELU Model: 4 Layers, 2048 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 2 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.9507. Min Act: -0.1700

Data Index: 29280 (C4 (Web Text))

Max Activating Token Index: 305

Text #1

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.9194. Min Act: -0.1700

Data Index: 1263693 (C4 (Web Text))

Max Activating Token Index: 42

Text #2

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.8882. Min Act: -0.1700

Data Index: 605114 (C4 (Web Text))

Max Activating Token Index: 176

Text #3

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7944. Min Act: -0.1700

Data Index: 269568 (C4 (Web Text))

Max Activating Token Index: 855

Text #4

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7319. Min Act: -0.1700

Data Index: 262703 (C4 (Web Text))

Max Activating Token Index: 1016

Text #5

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7944. Min Act: -0.1700

Data Index: 477012 (C4 (Web Text))

Max Activating Token Index: 708

Text #6

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7944. Min Act: -0.1700

Data Index: 653020 (C4 (Web Text))

Max Activating Token Index: 234

Text #7

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7007. Min Act: -0.1700

Data Index: 1337329 (C4 (Web Text))

Max Activating Token Index: 783

Text #8

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.7007. Min Act: -0.1700

Data Index: 174953 (C4 (Web Text))

Max Activating Token Index: 95

Text #9

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.6694. Min Act: -0.1700

Data Index: 193758 (C4 (Web Text))

Max Activating Token Index: 1002

Text #10

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.6382. Min Act: -0.1700

Data Index: 581843 (C4 (Web Text))

Max Activating Token Index: 30

Text #11

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.6382. Min Act: -0.1700

Data Index: 1025582 (C4 (Web Text))

Max Activating Token Index: 841

Text #12

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5444. Min Act: -0.1700

Data Index: 1167988 (C4 (Web Text))

Max Activating Token Index: 780

Text #13

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.6382. Min Act: -0.1700

Data Index: 686520 (C4 (Web Text))

Max Activating Token Index: 7

Text #14

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5444. Min Act: -0.1700

Data Index: 945530 (C4 (Web Text))

Max Activating Token Index: 688

Text #15

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5757. Min Act: -0.1700

Data Index: 171202 (C4 (Web Text))

Max Activating Token Index: 946

Text #16

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5757. Min Act: -0.1700

Data Index: 120695 (C4 (Web Text))

Max Activating Token Index: 1008

Text #17

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5757. Min Act: -0.1700

Data Index: 1092257 (C4 (Web Text))

Max Activating Token Index: 949

Text #18

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.6069. Min Act: -0.1700

Data Index: 815690 (C4 (Web Text))

Max Activating Token Index: 296

Text #19

Max Range: 5.9507. Min Range: -5.9507

Max Act: 5.5132. Min Act: -0.1700

Data Index: 809824 (C4 (Web Text))

Max Activating Token Index: 860