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

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

Neuron 1591 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: 0.0513. Min Range: -0.0513

Max Act: 0.0513. Min Act: -0.0003

Data Index: 1385995 (Python Code)

Max Activating Token Index: 477

Text #1

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0490. Min Act: -0.0003

Data Index: 368909 (C4 (Web Text))

Max Activating Token Index: 703

Text #2

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0490. Min Act: -0.0003

Data Index: 302900 (C4 (Web Text))

Max Activating Token Index: 123

Text #3

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0472. Min Act: -0.0003

Data Index: 232560 (C4 (Web Text))

Max Activating Token Index: 501

Text #4

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0450. Min Act: -0.0003

Data Index: 868660 (C4 (Web Text))

Max Activating Token Index: 718

Text #5

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0462. Min Act: -0.0003

Data Index: 667769 (C4 (Web Text))

Max Activating Token Index: 804

Text #6

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0452. Min Act: -0.0003

Data Index: 1279099 (C4 (Web Text))

Max Activating Token Index: 824

Text #7

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0443. Min Act: -0.0003

Data Index: 337683 (C4 (Web Text))

Max Activating Token Index: 758

Text #8

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0452. Min Act: -0.0003

Data Index: 1187871 (C4 (Web Text))

Max Activating Token Index: 197

Text #9

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0441. Min Act: -0.0003

Data Index: 273707 (C4 (Web Text))

Max Activating Token Index: 889

Text #10

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0441. Min Act: -0.0001

Data Index: 560637 (C4 (Web Text))

Max Activating Token Index: 814

Text #11

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0438. Min Act: -0.0003

Data Index: 2586 (C4 (Web Text))

Max Activating Token Index: 52

Text #12

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0427. Min Act: -0.0003

Data Index: 764792 (C4 (Web Text))

Max Activating Token Index: 603

Text #13

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0433. Min Act: -0.0002

Data Index: 118481 (C4 (Web Text))

Max Activating Token Index: 326

Text #14

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0434. Min Act: -0.0003

Data Index: 243292 (C4 (Web Text))

Max Activating Token Index: 117

Text #15

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0417. Min Act: -0.0003

Data Index: 406125 (C4 (Web Text))

Max Activating Token Index: 328

Text #16

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0416. Min Act: -0.0003

Data Index: 1086256 (C4 (Web Text))

Max Activating Token Index: 874

Text #17

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0415. Min Act: -0.0003

Data Index: 547388 (C4 (Web Text))

Max Activating Token Index: 32

Text #18

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0414. Min Act: -0.0003

Data Index: 443413 (C4 (Web Text))

Max Activating Token Index: 725

Text #19

Max Range: 0.0513. Min Range: -0.0513

Max Act: 0.0421. Min Act: -0.0002

Data Index: 1402970 (Python Code)

Max Activating Token Index: 87