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

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

Neuron 3005 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0189. Min Act: -0.0003

Data Index: 999949 (C4 (Web Text))

Max Activating Token Index: 967

Text #1

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0215. Min Act: -0.0003

Data Index: 450130 (C4 (Web Text))

Max Activating Token Index: 347

Text #2

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0180. Min Act: -0.0003

Data Index: 51606 (C4 (Web Text))

Max Activating Token Index: 296

Text #3

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0169. Min Act: -0.0003

Data Index: 823881 (C4 (Web Text))

Max Activating Token Index: 195

Text #4

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0182. Min Act: -0.0003

Data Index: 1338806 (C4 (Web Text))

Max Activating Token Index: 219

Text #5

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0191. Min Act: -0.0003

Data Index: 118675 (C4 (Web Text))

Max Activating Token Index: 142

Text #6

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0175. Min Act: -0.0003

Data Index: 795773 (C4 (Web Text))

Max Activating Token Index: 58

Text #7

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0185. Min Act: -0.0003

Data Index: 56674 (C4 (Web Text))

Max Activating Token Index: 315

Text #8

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0157. Min Act: -0.0003

Data Index: 563709 (C4 (Web Text))

Max Activating Token Index: 262

Text #9

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0159. Min Act: -0.0003

Data Index: 258534 (C4 (Web Text))

Max Activating Token Index: 925

Text #10

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0149. Min Act: -0.0003

Data Index: 1251095 (C4 (Web Text))

Max Activating Token Index: 792

Text #11

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0145. Min Act: -0.0003

Data Index: 1518194 (Python Code)

Max Activating Token Index: 676

Text #12

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0178. Min Act: -0.0003

Data Index: 1297240 (C4 (Web Text))

Max Activating Token Index: 92

Text #13

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0162. Min Act: -0.0003

Data Index: 604833 (C4 (Web Text))

Max Activating Token Index: 180

Text #14

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0160. Min Act: -0.0003

Data Index: 1648058 (Python Code)

Max Activating Token Index: 780

Text #15

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0166. Min Act: -0.0003

Data Index: 845070 (C4 (Web Text))

Max Activating Token Index: 289

Text #16

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0132. Min Act: -0.0003

Data Index: 807177 (C4 (Web Text))

Max Activating Token Index: 423

Text #17

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0153. Min Act: -0.0003

Data Index: 1576189 (Python Code)

Max Activating Token Index: 878

Text #18

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0138. Min Act: -0.0003

Data Index: 464229 (C4 (Web Text))

Max Activating Token Index: 802

Text #19

Max Range: 0.0215. Min Range: -0.0215

Max Act: 0.0157. Min Act: -0.0003

Data Index: 918688 (C4 (Web Text))

Max Activating Token Index: 580