Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 8 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0136. Min Act: -0.0006

Data Index: 356283 (C4 (Web Text))

Max Activating Token Index: 440

Text #1

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0129. Min Act: -0.0006

Data Index: 781988 (C4 (Web Text))

Max Activating Token Index: 354

Text #2

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0125. Min Act: -0.0006

Data Index: 857914 (C4 (Web Text))

Max Activating Token Index: 114

Text #3

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0127. Min Act: -0.0006

Data Index: 240604 (C4 (Web Text))

Max Activating Token Index: 196

Text #4

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0133. Min Act: -0.0006

Data Index: 1018489 (C4 (Web Text))

Max Activating Token Index: 895

Text #5

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0123. Min Act: -0.0006

Data Index: 67558 (C4 (Web Text))

Max Activating Token Index: 232

Text #6

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0119. Min Act: -0.0006

Data Index: 148651 (C4 (Web Text))

Max Activating Token Index: 357

Text #7

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0118. Min Act: -0.0006

Data Index: 490028 (C4 (Web Text))

Max Activating Token Index: 358

Text #8

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0120. Min Act: -0.0006

Data Index: 1233487 (C4 (Web Text))

Max Activating Token Index: 414

Text #9

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0114. Min Act: -0.0006

Data Index: 1647987 (Python Code)

Max Activating Token Index: 888

Text #10

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0117. Min Act: -0.0006

Data Index: 1186012 (C4 (Web Text))

Max Activating Token Index: 55

Text #11

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0117. Min Act: -0.0006

Data Index: 805252 (C4 (Web Text))

Max Activating Token Index: 437

Text #12

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0120. Min Act: -0.0006

Data Index: 348282 (C4 (Web Text))

Max Activating Token Index: 666

Text #13

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0119. Min Act: -0.0006

Data Index: 469959 (C4 (Web Text))

Max Activating Token Index: 315

Text #14

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0115. Min Act: -0.0006

Data Index: 313318 (C4 (Web Text))

Max Activating Token Index: 766

Text #15

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0112. Min Act: -0.0006

Data Index: 1599304 (Python Code)

Max Activating Token Index: 614

Text #16

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0114. Min Act: -0.0006

Data Index: 124606 (C4 (Web Text))

Max Activating Token Index: 165

Text #17

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0117. Min Act: -0.0006

Data Index: 1209707 (C4 (Web Text))

Max Activating Token Index: 78

Text #18

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0114. Min Act: -0.0006

Data Index: 246160 (C4 (Web Text))

Max Activating Token Index: 598

Text #19

Max Range: 0.0136. Min Range: -0.0136

Max Act: 0.0111. Min Act: -0.0006

Data Index: 62768 (C4 (Web Text))

Max Activating Token Index: 953