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

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

Neuron 655 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: 0.0283. Min Range: -0.0283

Max Act: 0.0283. Min Act: -0.0005

Data Index: 83349 (C4 (Web Text))

Max Activating Token Index: 431

Text #1

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0258. Min Act: -0.0005

Data Index: 457754 (C4 (Web Text))

Max Activating Token Index: 688

Text #2

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0233. Min Act: -0.0006

Data Index: 477075 (C4 (Web Text))

Max Activating Token Index: 89

Text #3

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0232. Min Act: -0.0006

Data Index: 920924 (C4 (Web Text))

Max Activating Token Index: 419

Text #4

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0224. Min Act: -0.0006

Data Index: 592395 (C4 (Web Text))

Max Activating Token Index: 806

Text #5

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0218. Min Act: -0.0006

Data Index: 474772 (C4 (Web Text))

Max Activating Token Index: 315

Text #6

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0230. Min Act: -0.0006

Data Index: 1249702 (C4 (Web Text))

Max Activating Token Index: 584

Text #7

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0221. Min Act: -0.0006

Data Index: 561838 (C4 (Web Text))

Max Activating Token Index: 701

Text #8

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0226. Min Act: -0.0005

Data Index: 656237 (C4 (Web Text))

Max Activating Token Index: 253

Text #9

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0214. Min Act: -0.0005

Data Index: 262896 (C4 (Web Text))

Max Activating Token Index: 478

Text #10

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0213. Min Act: -0.0006

Data Index: 941750 (C4 (Web Text))

Max Activating Token Index: 555

Text #11

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0212. Min Act: -0.0006

Data Index: 425629 (C4 (Web Text))

Max Activating Token Index: 308

Text #12

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0211. Min Act: -0.0006

Data Index: 123700 (C4 (Web Text))

Max Activating Token Index: 223

Text #13

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0221. Min Act: -0.0006

Data Index: 926941 (C4 (Web Text))

Max Activating Token Index: 532

Text #14

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0212. Min Act: -0.0006

Data Index: 1016347 (C4 (Web Text))

Max Activating Token Index: 469

Text #15

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0206. Min Act: -0.0006

Data Index: 846409 (C4 (Web Text))

Max Activating Token Index: 251

Text #16

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0214. Min Act: -0.0005

Data Index: 1058291 (C4 (Web Text))

Max Activating Token Index: 53

Text #17

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0210. Min Act: -0.0006

Data Index: 999763 (C4 (Web Text))

Max Activating Token Index: 840

Text #18

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0208. Min Act: -0.0006

Data Index: 1216305 (C4 (Web Text))

Max Activating Token Index: 645

Text #19

Max Range: 0.0283. Min Range: -0.0283

Max Act: 0.0207. Min Act: -0.0006

Data Index: 128908 (C4 (Web Text))

Max Activating Token Index: 276