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

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

Neuron 202 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: 0.0193. Min Range: -0.0193

Max Act: 0.0188. Min Act: -0.0006

Data Index: 1517414 (Python Code)

Max Activating Token Index: 905

Text #1

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0193. Min Act: -0.0006

Data Index: 582473 (C4 (Web Text))

Max Activating Token Index: 207

Text #2

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0178. Min Act: -0.0006

Data Index: 906508 (C4 (Web Text))

Max Activating Token Index: 79

Text #3

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0178. Min Act: -0.0006

Data Index: 722881 (C4 (Web Text))

Max Activating Token Index: 722

Text #4

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0187. Min Act: -0.0006

Data Index: 445167 (C4 (Web Text))

Max Activating Token Index: 234

Text #5

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0166. Min Act: -0.0006

Data Index: 1080012 (C4 (Web Text))

Max Activating Token Index: 925

Text #6

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0168. Min Act: -0.0006

Data Index: 784931 (C4 (Web Text))

Max Activating Token Index: 889

Text #7

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0175. Min Act: -0.0006

Data Index: 661224 (C4 (Web Text))

Max Activating Token Index: 345

Text #8

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0179. Min Act: -0.0006

Data Index: 496384 (C4 (Web Text))

Max Activating Token Index: 220

Text #9

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0170. Min Act: -0.0006

Data Index: 303591 (C4 (Web Text))

Max Activating Token Index: 553

Text #10

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0176. Min Act: -0.0006

Data Index: 894087 (C4 (Web Text))

Max Activating Token Index: 775

Text #11

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0153. Min Act: -0.0006

Data Index: 292740 (C4 (Web Text))

Max Activating Token Index: 262

Text #12

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0183. Min Act: -0.0006

Data Index: 997882 (C4 (Web Text))

Max Activating Token Index: 783

Text #13

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0163. Min Act: -0.0006

Data Index: 1472797 (Python Code)

Max Activating Token Index: 616

Text #14

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0166. Min Act: -0.0006

Data Index: 675444 (C4 (Web Text))

Max Activating Token Index: 179

Text #15

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0173. Min Act: -0.0006

Data Index: 468620 (C4 (Web Text))

Max Activating Token Index: 176

Text #16

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0161. Min Act: -0.0006

Data Index: 414698 (C4 (Web Text))

Max Activating Token Index: 844

Text #17

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0167. Min Act: -0.0005

Data Index: 77180 (C4 (Web Text))

Max Activating Token Index: 247

Text #18

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0167. Min Act: -0.0006

Data Index: 106402 (C4 (Web Text))

Max Activating Token Index: 18

Text #19

Max Range: 0.0193. Min Range: -0.0193

Max Act: 0.0162. Min Act: -0.0006

Data Index: 237831 (C4 (Web Text))

Max Activating Token Index: 694