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

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

Neuron 392 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0172. Min Act: -0.0006

Data Index: 1097269 (C4 (Web Text))

Max Activating Token Index: 735

Text #1

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0156. Min Act: -0.0006

Data Index: 722482 (C4 (Web Text))

Max Activating Token Index: 140

Text #2

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0156. Min Act: -0.0006

Data Index: 537990 (C4 (Web Text))

Max Activating Token Index: 563

Text #3

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0146. Min Act: -0.0006

Data Index: 39512 (C4 (Web Text))

Max Activating Token Index: 597

Text #4

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0150. Min Act: -0.0006

Data Index: 574984 (C4 (Web Text))

Max Activating Token Index: 519

Text #5

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0142. Min Act: -0.0006

Data Index: 1282844 (C4 (Web Text))

Max Activating Token Index: 87

Text #6

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0139. Min Act: -0.0006

Data Index: 9335 (C4 (Web Text))

Max Activating Token Index: 922

Text #7

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0142. Min Act: -0.0006

Data Index: 1086712 (C4 (Web Text))

Max Activating Token Index: 412

Text #8

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0138. Min Act: -0.0006

Data Index: 423037 (C4 (Web Text))

Max Activating Token Index: 409

Text #9

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0137. Min Act: -0.0006

Data Index: 20272 (C4 (Web Text))

Max Activating Token Index: 1018

Text #10

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0137. Min Act: -0.0005

Data Index: 1620556 (Python Code)

Max Activating Token Index: 658

Text #11

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0138. Min Act: -0.0006

Data Index: 908610 (C4 (Web Text))

Max Activating Token Index: 601

Text #12

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0133. Min Act: -0.0006

Data Index: 434027 (C4 (Web Text))

Max Activating Token Index: 790

Text #13

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0132. Min Act: -0.0005

Data Index: 334037 (C4 (Web Text))

Max Activating Token Index: 753

Text #14

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0132. Min Act: -0.0006

Data Index: 1436104 (Python Code)

Max Activating Token Index: 591

Text #15

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0134. Min Act: -0.0006

Data Index: 428232 (C4 (Web Text))

Max Activating Token Index: 96

Text #16

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0137. Min Act: -0.0006

Data Index: 976836 (C4 (Web Text))

Max Activating Token Index: 200

Text #17

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0134. Min Act: -0.0006

Data Index: 612053 (C4 (Web Text))

Max Activating Token Index: 829

Text #18

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0133. Min Act: -0.0006

Data Index: 991844 (C4 (Web Text))

Max Activating Token Index: 747

Text #19

Max Range: 0.0172. Min Range: -0.0172

Max Act: 0.0130. Min Act: -0.0006

Data Index: 1126678 (C4 (Web Text))

Max Activating Token Index: 785