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

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

Neuron 1000 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: 0.0199. Min Range: -0.0199

Max Act: 0.0199. Min Act: -0.0001

Data Index: 587735 (C4 (Web Text))

Max Activating Token Index: 641

Text #1

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0172. Min Act: -0.0001

Data Index: 200812 (C4 (Web Text))

Max Activating Token Index: 534

Text #2

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0170. Min Act: -0.0001

Data Index: 829440 (C4 (Web Text))

Max Activating Token Index: 690

Text #3

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0163. Min Act: -0.0001

Data Index: 910300 (C4 (Web Text))

Max Activating Token Index: 116

Text #4

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0155. Min Act: -0.0001

Data Index: 1455472 (Python Code)

Max Activating Token Index: 324

Text #5

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0151. Min Act: -0.0001

Data Index: 1548708 (Python Code)

Max Activating Token Index: 205

Text #6

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0139. Min Act: -0.0001

Data Index: 1340300 (C4 (Web Text))

Max Activating Token Index: 704

Text #7

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0153. Min Act: -0.0001

Data Index: 1353008 (C4 (Web Text))

Max Activating Token Index: 541

Text #8

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0144. Min Act: -0.0001

Data Index: 845585 (C4 (Web Text))

Max Activating Token Index: 502

Text #9

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0138. Min Act: -0.0001

Data Index: 1603590 (Python Code)

Max Activating Token Index: 727

Text #10

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0147. Min Act: -0.0001

Data Index: 723143 (C4 (Web Text))

Max Activating Token Index: 477

Text #11

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0148. Min Act: -0.0001

Data Index: 1596999 (Python Code)

Max Activating Token Index: 495

Text #12

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0140. Min Act: -0.0001

Data Index: 181948 (C4 (Web Text))

Max Activating Token Index: 181

Text #13

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0142. Min Act: -0.0001

Data Index: 888674 (C4 (Web Text))

Max Activating Token Index: 732

Text #14

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0133. Min Act: -0.0001

Data Index: 1445384 (Python Code)

Max Activating Token Index: 184

Text #15

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0136. Min Act: -0.0001

Data Index: 923004 (C4 (Web Text))

Max Activating Token Index: 302

Text #16

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0142. Min Act: -0.0001

Data Index: 849329 (C4 (Web Text))

Max Activating Token Index: 420

Text #17

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0138. Min Act: -0.0001

Data Index: 842902 (C4 (Web Text))

Max Activating Token Index: 801

Text #18

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0141. Min Act: -0.0001

Data Index: 346086 (C4 (Web Text))

Max Activating Token Index: 242

Text #19

Max Range: 0.0199. Min Range: -0.0199

Max Act: 0.0138. Min Act: -0.0001

Data Index: 715785 (C4 (Web Text))

Max Activating Token Index: 585