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

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

Neuron 1902 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: 0.0430. Min Range: -0.0430

Max Act: 0.0427. Min Act: -0.0003

Data Index: 1284199 (C4 (Web Text))

Max Activating Token Index: 63

Text #1

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0430. Min Act: -0.0003

Data Index: 1003658 (C4 (Web Text))

Max Activating Token Index: 276

Text #2

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0427. Min Act: -0.0003

Data Index: 1350299 (C4 (Web Text))

Max Activating Token Index: 783

Text #3

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0404. Min Act: -0.0004

Data Index: 808147 (C4 (Web Text))

Max Activating Token Index: 963

Text #4

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0381. Min Act: -0.0003

Data Index: 811703 (C4 (Web Text))

Max Activating Token Index: 533

Text #5

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0392. Min Act: -0.0003

Data Index: 403285 (C4 (Web Text))

Max Activating Token Index: 815

Text #6

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0387. Min Act: -0.0003

Data Index: 127011 (C4 (Web Text))

Max Activating Token Index: 476

Text #7

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0379. Min Act: -0.0003

Data Index: 759506 (C4 (Web Text))

Max Activating Token Index: 540

Text #8

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0377. Min Act: -0.0003

Data Index: 59442 (C4 (Web Text))

Max Activating Token Index: 760

Text #9

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0391. Min Act: -0.0003

Data Index: 652484 (C4 (Web Text))

Max Activating Token Index: 921

Text #10

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0397. Min Act: -0.0003

Data Index: 582331 (C4 (Web Text))

Max Activating Token Index: 66

Text #11

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0375. Min Act: -0.0003

Data Index: 959943 (C4 (Web Text))

Max Activating Token Index: 191

Text #12

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0391. Min Act: -0.0003

Data Index: 1545460 (Python Code)

Max Activating Token Index: 695

Text #13

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0396. Min Act: -0.0003

Data Index: 654108 (C4 (Web Text))

Max Activating Token Index: 149

Text #14

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0376. Min Act: -0.0003

Data Index: 1487075 (Python Code)

Max Activating Token Index: 905

Text #15

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0372. Min Act: -0.0003

Data Index: 1605827 (Python Code)

Max Activating Token Index: 111

Text #16

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0377. Min Act: -0.0003

Data Index: 1077858 (C4 (Web Text))

Max Activating Token Index: 430

Text #17

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0353. Min Act: -0.0003

Data Index: 762486 (C4 (Web Text))

Max Activating Token Index: 330

Text #18

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0367. Min Act: -0.0002

Data Index: 1128978 (C4 (Web Text))

Max Activating Token Index: 427

Text #19

Max Range: 0.0430. Min Range: -0.0430

Max Act: 0.0361. Min Act: -0.0004

Data Index: 426358 (C4 (Web Text))

Max Activating Token Index: 262