Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 1593 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.4774. Min Act: -0.0000

Data Index: 1436240 (Python Code)

Max Activating Token Index: 233

Text #1

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.5887. Min Act: -0.0000

Data Index: 1238294 (C4 (Web Text))

Max Activating Token Index: 71

Text #2

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.3941. Min Act: -0.0000

Data Index: 1438468 (Python Code)

Max Activating Token Index: 603

Text #3

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.2264. Min Act: -0.0000

Data Index: 940402 (C4 (Web Text))

Max Activating Token Index: 370

Text #4

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.4933. Min Act: -0.0000

Data Index: 1278428 (C4 (Web Text))

Max Activating Token Index: 661

Text #5

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.0065. Min Act: -0.0000

Data Index: 1436191 (Python Code)

Max Activating Token Index: 285

Text #6

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.0620. Min Act: -0.0000

Data Index: 843315 (C4 (Web Text))

Max Activating Token Index: 729

Text #7

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.2620. Min Act: -0.0000

Data Index: 851091 (C4 (Web Text))

Max Activating Token Index: 825

Text #8

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.8114. Min Act: -0.0000

Data Index: 1001666 (C4 (Web Text))

Max Activating Token Index: 94

Text #9

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.1083. Min Act: -0.0000

Data Index: 1645631 (Python Code)

Max Activating Token Index: 573

Text #10

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.9399. Min Act: -0.0000

Data Index: 388108 (C4 (Web Text))

Max Activating Token Index: 91

Text #11

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.9215. Min Act: -0.0000

Data Index: 56998 (C4 (Web Text))

Max Activating Token Index: 877

Text #12

Max Range: 3.5887. Min Range: -3.5887

Max Act: 3.0588. Min Act: -0.0000

Data Index: 1281508 (C4 (Web Text))

Max Activating Token Index: 918

Text #13

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.8774. Min Act: -0.0000

Data Index: 1470862 (Python Code)

Max Activating Token Index: 494

Text #14

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.7103. Min Act: -0.0000

Data Index: 896919 (C4 (Web Text))

Max Activating Token Index: 297

Text #15

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.8483. Min Act: -0.0000

Data Index: 1564862 (Python Code)

Max Activating Token Index: 806

Text #16

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.8116. Min Act: -0.0000

Data Index: 1563037 (Python Code)

Max Activating Token Index: 157

Text #17

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.7766. Min Act: -0.0000

Data Index: 1433256 (Python Code)

Max Activating Token Index: 856

Text #18

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.6504. Min Act: -0.0000

Data Index: 1507079 (Python Code)

Max Activating Token Index: 725

Text #19

Max Range: 3.5887. Min Range: -3.5887

Max Act: 2.7577. Min Act: -0.0000

Data Index: 1390894 (Python Code)

Max Activating Token Index: 16