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

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

Neuron 1638 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0345. Min Act: -0.0001

Data Index: 758770 (C4 (Web Text))

Max Activating Token Index: 582

Text #1

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0293. Min Act: -0.0002

Data Index: 615003 (C4 (Web Text))

Max Activating Token Index: 773

Text #2

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0304. Min Act: -0.0002

Data Index: 329635 (C4 (Web Text))

Max Activating Token Index: 355

Text #3

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0296. Min Act: -0.0002

Data Index: 105007 (C4 (Web Text))

Max Activating Token Index: 894

Text #4

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0291. Min Act: -0.0002

Data Index: 1621867 (Python Code)

Max Activating Token Index: 960

Text #5

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0287. Min Act: -0.0001

Data Index: 1013390 (C4 (Web Text))

Max Activating Token Index: 876

Text #6

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0276. Min Act: -0.0003

Data Index: 265293 (C4 (Web Text))

Max Activating Token Index: 496

Text #7

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0281. Min Act: -0.0002

Data Index: 1255868 (C4 (Web Text))

Max Activating Token Index: 748

Text #8

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0283. Min Act: -0.0002

Data Index: 820454 (C4 (Web Text))

Max Activating Token Index: 181

Text #9

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0271. Min Act: -0.0002

Data Index: 329882 (C4 (Web Text))

Max Activating Token Index: 421

Text #10

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0275. Min Act: -0.0003

Data Index: 336567 (C4 (Web Text))

Max Activating Token Index: 390

Text #11

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0281. Min Act: -0.0003

Data Index: 716726 (C4 (Web Text))

Max Activating Token Index: 592

Text #12

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0277. Min Act: -0.0002

Data Index: 767686 (C4 (Web Text))

Max Activating Token Index: 895

Text #13

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0271. Min Act: -0.0002

Data Index: 700882 (C4 (Web Text))

Max Activating Token Index: 694

Text #14

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0277. Min Act: -0.0002

Data Index: 836708 (C4 (Web Text))

Max Activating Token Index: 287

Text #15

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0270. Min Act: -0.0002

Data Index: 279203 (C4 (Web Text))

Max Activating Token Index: 349

Text #16

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0267. Min Act: -0.0002

Data Index: 85845 (C4 (Web Text))

Max Activating Token Index: 942

Text #17

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0270. Min Act: -0.0002

Data Index: 1080844 (C4 (Web Text))

Max Activating Token Index: 952

Text #18

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0275. Min Act: -0.0002

Data Index: 370269 (C4 (Web Text))

Max Activating Token Index: 469

Text #19

Max Range: 0.0345. Min Range: -0.0345

Max Act: 0.0268. Min Act: -0.0002

Data Index: 1442776 (Python Code)

Max Activating Token Index: 950