Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 848 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0249. Min Act: -0.0000

Data Index: 495837 (C4 (Web Text))

Max Activating Token Index: 948

Text #1

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0220. Min Act: -0.0000

Data Index: 473107 (C4 (Web Text))

Max Activating Token Index: 606

Text #2

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0237. Min Act: -0.0000

Data Index: 689445 (C4 (Web Text))

Max Activating Token Index: 986

Text #3

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0227. Min Act: -0.0000

Data Index: 466616 (C4 (Web Text))

Max Activating Token Index: 315

Text #4

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0223. Min Act: -0.0000

Data Index: 301035 (C4 (Web Text))

Max Activating Token Index: 992

Text #5

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0204. Min Act: -0.0000

Data Index: 1548977 (Python Code)

Max Activating Token Index: 694

Text #6

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0222. Min Act: -0.0000

Data Index: 252584 (C4 (Web Text))

Max Activating Token Index: 269

Text #7

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0217. Min Act: -0.0000

Data Index: 495847 (C4 (Web Text))

Max Activating Token Index: 1008

Text #8

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0218. Min Act: -0.0000

Data Index: 356602 (C4 (Web Text))

Max Activating Token Index: 378

Text #9

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0207. Min Act: -0.0000

Data Index: 1512091 (Python Code)

Max Activating Token Index: 289

Text #10

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0214. Min Act: -0.0000

Data Index: 4882 (C4 (Web Text))

Max Activating Token Index: 135

Text #11

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0203. Min Act: -0.0000

Data Index: 1377523 (Python Code)

Max Activating Token Index: 925

Text #12

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0201. Min Act: -0.0000

Data Index: 1291681 (C4 (Web Text))

Max Activating Token Index: 131

Text #13

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0197. Min Act: -0.0000

Data Index: 1413955 (Python Code)

Max Activating Token Index: 430

Text #14

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0218. Min Act: -0.0000

Data Index: 217807 (C4 (Web Text))

Max Activating Token Index: 833

Text #15

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0196. Min Act: -0.0000

Data Index: 569156 (C4 (Web Text))

Max Activating Token Index: 820

Text #16

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0198. Min Act: -0.0000

Data Index: 1312234 (C4 (Web Text))

Max Activating Token Index: 800

Text #17

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0192. Min Act: -0.0000

Data Index: 622595 (C4 (Web Text))

Max Activating Token Index: 221

Text #18

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0208. Min Act: -0.0000

Data Index: 10237 (C4 (Web Text))

Max Activating Token Index: 737

Text #19

Max Range: 0.0249. Min Range: -0.0249

Max Act: 0.0200. Min Act: -0.0000

Data Index: 1621376 (Python Code)

Max Activating Token Index: 35