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

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

Neuron 501 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: 0.0511. Min Range: -0.0511

Max Act: 0.0511. Min Act: -0.0006

Data Index: 1290593 (C4 (Web Text))

Max Activating Token Index: 1002

Text #1

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0461. Min Act: -0.0006

Data Index: 1193103 (C4 (Web Text))

Max Activating Token Index: 560

Text #2

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0398. Min Act: -0.0006

Data Index: 535619 (C4 (Web Text))

Max Activating Token Index: 544

Text #3

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0414. Min Act: -0.0006

Data Index: 796521 (C4 (Web Text))

Max Activating Token Index: 149

Text #4

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0393. Min Act: -0.0006

Data Index: 559190 (C4 (Web Text))

Max Activating Token Index: 528

Text #5

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0418. Min Act: -0.0006

Data Index: 896651 (C4 (Web Text))

Max Activating Token Index: 329

Text #6

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0397. Min Act: -0.0006

Data Index: 197051 (C4 (Web Text))

Max Activating Token Index: 492

Text #7

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0378. Min Act: -0.0005

Data Index: 1338258 (C4 (Web Text))

Max Activating Token Index: 440

Text #8

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0440. Min Act: -0.0006

Data Index: 142842 (C4 (Web Text))

Max Activating Token Index: 609

Text #9

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0420. Min Act: -0.0005

Data Index: 759591 (C4 (Web Text))

Max Activating Token Index: 612

Text #10

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0367. Min Act: -0.0006

Data Index: 818245 (C4 (Web Text))

Max Activating Token Index: 462

Text #11

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0399. Min Act: -0.0006

Data Index: 1286174 (C4 (Web Text))

Max Activating Token Index: 471

Text #12

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0376. Min Act: -0.0006

Data Index: 39600 (C4 (Web Text))

Max Activating Token Index: 583

Text #13

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0386. Min Act: -0.0006

Data Index: 1207509 (C4 (Web Text))

Max Activating Token Index: 877

Text #14

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0381. Min Act: -0.0006

Data Index: 942896 (C4 (Web Text))

Max Activating Token Index: 627

Text #15

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0379. Min Act: -0.0005

Data Index: 137294 (C4 (Web Text))

Max Activating Token Index: 675

Text #16

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0385. Min Act: -0.0006

Data Index: 1606819 (Python Code)

Max Activating Token Index: 746

Text #17

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0374. Min Act: -0.0006

Data Index: 955428 (C4 (Web Text))

Max Activating Token Index: 118

Text #18

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0385. Min Act: -0.0006

Data Index: 785777 (C4 (Web Text))

Max Activating Token Index: 234

Text #19

Max Range: 0.0511. Min Range: -0.0511

Max Act: 0.0388. Min Act: -0.0006

Data Index: 564266 (C4 (Web Text))

Max Activating Token Index: 973