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

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

Neuron 1973 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: 0.0473. Min Range: -0.0473

Max Act: 0.0473. Min Act: -0.0003

Data Index: 333552 (C4 (Web Text))

Max Activating Token Index: 621

Text #1

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0452. Min Act: -0.0002

Data Index: 941758 (C4 (Web Text))

Max Activating Token Index: 467

Text #2

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0436. Min Act: -0.0002

Data Index: 227125 (C4 (Web Text))

Max Activating Token Index: 210

Text #3

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0445. Min Act: -0.0003

Data Index: 59056 (C4 (Web Text))

Max Activating Token Index: 127

Text #4

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0427. Min Act: -0.0003

Data Index: 985799 (C4 (Web Text))

Max Activating Token Index: 246

Text #5

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0423. Min Act: -0.0003

Data Index: 233212 (C4 (Web Text))

Max Activating Token Index: 233

Text #6

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0407. Min Act: -0.0004

Data Index: 1067744 (C4 (Web Text))

Max Activating Token Index: 654

Text #7

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0398. Min Act: -0.0002

Data Index: 45341 (C4 (Web Text))

Max Activating Token Index: 889

Text #8

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0395. Min Act: -0.0003

Data Index: 179503 (C4 (Web Text))

Max Activating Token Index: 620

Text #9

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0395. Min Act: -0.0003

Data Index: 1280072 (C4 (Web Text))

Max Activating Token Index: 503

Text #10

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0388. Min Act: -0.0003

Data Index: 1398076 (Python Code)

Max Activating Token Index: 474

Text #11

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0399. Min Act: -0.0003

Data Index: 904752 (C4 (Web Text))

Max Activating Token Index: 675

Text #12

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0391. Min Act: -0.0003

Data Index: 976296 (C4 (Web Text))

Max Activating Token Index: 126

Text #13

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0376. Min Act: -0.0003

Data Index: 1199961 (C4 (Web Text))

Max Activating Token Index: 628

Text #14

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0386. Min Act: -0.0003

Data Index: 1120073 (C4 (Web Text))

Max Activating Token Index: 629

Text #15

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0387. Min Act: -0.0003

Data Index: 877634 (C4 (Web Text))

Max Activating Token Index: 405

Text #16

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0389. Min Act: -0.0003

Data Index: 1608670 (Python Code)

Max Activating Token Index: 764

Text #17

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0388. Min Act: -0.0003

Data Index: 1527341 (Python Code)

Max Activating Token Index: 498

Text #18

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0381. Min Act: -0.0003

Data Index: 221647 (C4 (Web Text))

Max Activating Token Index: 153

Text #19

Max Range: 0.0473. Min Range: -0.0473

Max Act: 0.0378. Min Act: -0.0003

Data Index: 953555 (C4 (Web Text))

Max Activating Token Index: 348