Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 340 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0359. Min Act: -0.0006

Data Index: 666228 (C4 (Web Text))

Max Activating Token Index: 1012

Text #1

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0292. Min Act: -0.0006

Data Index: 415898 (C4 (Web Text))

Max Activating Token Index: 1

Text #2

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0286. Min Act: -0.0006

Data Index: 478372 (C4 (Web Text))

Max Activating Token Index: 229

Text #3

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0290. Min Act: -0.0005

Data Index: 1450490 (Python Code)

Max Activating Token Index: 194

Text #4

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0275. Min Act: -0.0006

Data Index: 320540 (C4 (Web Text))

Max Activating Token Index: 519

Text #5

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0264. Min Act: -0.0006

Data Index: 760887 (C4 (Web Text))

Max Activating Token Index: 625

Text #6

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0277. Min Act: -0.0006

Data Index: 1298371 (C4 (Web Text))

Max Activating Token Index: 570

Text #7

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0268. Min Act: -0.0006

Data Index: 1213489 (C4 (Web Text))

Max Activating Token Index: 444

Text #8

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0263. Min Act: -0.0006

Data Index: 810902 (C4 (Web Text))

Max Activating Token Index: 75

Text #9

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0269. Min Act: -0.0006

Data Index: 1272780 (C4 (Web Text))

Max Activating Token Index: 548

Text #10

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0269. Min Act: -0.0006

Data Index: 358224 (C4 (Web Text))

Max Activating Token Index: 1

Text #11

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0269. Min Act: -0.0006

Data Index: 567708 (C4 (Web Text))

Max Activating Token Index: 1

Text #12

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0265. Min Act: -0.0006

Data Index: 401365 (C4 (Web Text))

Max Activating Token Index: 529

Text #13

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0254. Min Act: -0.0006

Data Index: 574376 (C4 (Web Text))

Max Activating Token Index: 904

Text #14

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0247. Min Act: -0.0006

Data Index: 426966 (C4 (Web Text))

Max Activating Token Index: 1007

Text #15

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0252. Min Act: -0.0006

Data Index: 326306 (C4 (Web Text))

Max Activating Token Index: 644

Text #16

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0250. Min Act: -0.0006

Data Index: 426286 (C4 (Web Text))

Max Activating Token Index: 903

Text #17

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0248. Min Act: -0.0005

Data Index: 1079928 (C4 (Web Text))

Max Activating Token Index: 1

Text #18

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0248. Min Act: -0.0006

Data Index: 1604510 (Python Code)

Max Activating Token Index: 1

Text #19

Max Range: 0.0359. Min Range: -0.0359

Max Act: 0.0248. Min Act: -0.0006

Data Index: 791943 (C4 (Web Text))

Max Activating Token Index: 1018