Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3220 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1451. Min Act: -0.0001

Data Index: 691363 (C4 (Web Text))

Max Activating Token Index: 365

Text #1

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1495. Min Act: -0.0001

Data Index: 1173385 (C4 (Web Text))

Max Activating Token Index: 152

Text #2

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1386. Min Act: -0.0001

Data Index: 536065 (C4 (Web Text))

Max Activating Token Index: 151

Text #3

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1328. Min Act: -0.0001

Data Index: 31123 (C4 (Web Text))

Max Activating Token Index: 1022

Text #4

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1373. Min Act: -0.0001

Data Index: 964525 (C4 (Web Text))

Max Activating Token Index: 617

Text #5

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1298. Min Act: -0.0001

Data Index: 488523 (C4 (Web Text))

Max Activating Token Index: 545

Text #6

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1348. Min Act: -0.0001

Data Index: 362592 (C4 (Web Text))

Max Activating Token Index: 722

Text #7

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1327. Min Act: -0.0001

Data Index: 310017 (C4 (Web Text))

Max Activating Token Index: 744

Text #8

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1315. Min Act: -0.0001

Data Index: 172468 (C4 (Web Text))

Max Activating Token Index: 231

Text #9

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1138. Min Act: -0.0001

Data Index: 1634218 (Python Code)

Max Activating Token Index: 391

Text #10

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1254. Min Act: -0.0001

Data Index: 263457 (C4 (Web Text))

Max Activating Token Index: 644

Text #11

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1099. Min Act: -0.0001

Data Index: 1188123 (C4 (Web Text))

Max Activating Token Index: 478

Text #12

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1100. Min Act: -0.0001

Data Index: 195435 (C4 (Web Text))

Max Activating Token Index: 639

Text #13

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1158. Min Act: -0.0001

Data Index: 1007811 (C4 (Web Text))

Max Activating Token Index: 687

Text #14

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1112. Min Act: -0.0001

Data Index: 983194 (C4 (Web Text))

Max Activating Token Index: 851

Text #15

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.0948. Min Act: -0.0001

Data Index: 29918 (C4 (Web Text))

Max Activating Token Index: 862

Text #16

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1061. Min Act: -0.0001

Data Index: 1081981 (C4 (Web Text))

Max Activating Token Index: 1003

Text #17

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1054. Min Act: -0.0001

Data Index: 292389 (C4 (Web Text))

Max Activating Token Index: 536

Text #18

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1052. Min Act: -0.0001

Data Index: 29579 (C4 (Web Text))

Max Activating Token Index: 430

Text #19

Max Range: 0.1495. Min Range: -0.1495

Max Act: 0.1090. Min Act: -0.0001

Data Index: 186602 (C4 (Web Text))

Max Activating Token Index: 977