Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 3227 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 3.4731. Min Range: -3.4731

Max Act: 3.4731. Min Act: -0.0000

Data Index: 95675 (C4 (Web Text))

Max Activating Token Index: 911

Text #1

Max Range: 3.4731. Min Range: -3.4731

Max Act: 3.0698. Min Act: -0.0000

Data Index: 1531097 (Python Code)

Max Activating Token Index: 648

Text #2

Max Range: 3.4731. Min Range: -3.4731

Max Act: 3.1108. Min Act: -0.0000

Data Index: 190088 (C4 (Web Text))

Max Activating Token Index: 184

Text #3

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.6450. Min Act: -0.0000

Data Index: 1167235 (C4 (Web Text))

Max Activating Token Index: 928

Text #4

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.9383. Min Act: -0.0000

Data Index: 1507441 (Python Code)

Max Activating Token Index: 245

Text #5

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.9236. Min Act: -0.0000

Data Index: 1201933 (C4 (Web Text))

Max Activating Token Index: 982

Text #6

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.7294. Min Act: -0.0000

Data Index: 1178950 (C4 (Web Text))

Max Activating Token Index: 98

Text #7

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.8037. Min Act: -0.0000

Data Index: 742552 (C4 (Web Text))

Max Activating Token Index: 397

Text #8

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.7750. Min Act: -0.0000

Data Index: 378432 (C4 (Web Text))

Max Activating Token Index: 883

Text #9

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.6102. Min Act: -0.0000

Data Index: 807325 (C4 (Web Text))

Max Activating Token Index: 111

Text #10

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.7317. Min Act: -0.0000

Data Index: 135840 (C4 (Web Text))

Max Activating Token Index: 450

Text #11

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.5549. Min Act: -0.0000

Data Index: 129767 (C4 (Web Text))

Max Activating Token Index: 923

Text #12

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.7222. Min Act: -0.0000

Data Index: 1545578 (Python Code)

Max Activating Token Index: 824

Text #13

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.5422. Min Act: -0.0000

Data Index: 1188211 (C4 (Web Text))

Max Activating Token Index: 438

Text #14

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.6883. Min Act: -0.0000

Data Index: 665912 (C4 (Web Text))

Max Activating Token Index: 394

Text #15

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.5090. Min Act: -0.0000

Data Index: 883032 (C4 (Web Text))

Max Activating Token Index: 65

Text #16

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.6154. Min Act: -0.0000

Data Index: 596045 (C4 (Web Text))

Max Activating Token Index: 1014

Text #17

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.5974. Min Act: -0.0000

Data Index: 772568 (C4 (Web Text))

Max Activating Token Index: 153

Text #18

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.5665. Min Act: -0.0000

Data Index: 530087 (C4 (Web Text))

Max Activating Token Index: 260

Text #19

Max Range: 3.4731. Min Range: -3.4731

Max Act: 2.4875. Min Act: -0.0000

Data Index: 1333059 (C4 (Web Text))

Max Activating Token Index: 178