Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 405 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.4120. Min Act: -0.0000

Data Index: 1630662 (Python Code)

Max Activating Token Index: 971

Text #1

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3513. Min Act: -0.0000

Data Index: 1329368 (C4 (Web Text))

Max Activating Token Index: 358

Text #2

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3750. Min Act: -0.0000

Data Index: 1001119 (C4 (Web Text))

Max Activating Token Index: 838

Text #3

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3481. Min Act: -0.0000

Data Index: 443107 (C4 (Web Text))

Max Activating Token Index: 488

Text #4

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3465. Min Act: -0.0000

Data Index: 1458652 (Python Code)

Max Activating Token Index: 472

Text #5

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3355. Min Act: -0.0000

Data Index: 232596 (C4 (Web Text))

Max Activating Token Index: 868

Text #6

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3608. Min Act: -0.0000

Data Index: 124922 (C4 (Web Text))

Max Activating Token Index: 233

Text #7

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3496. Min Act: -0.0000

Data Index: 1065858 (C4 (Web Text))

Max Activating Token Index: 775

Text #8

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3164. Min Act: -0.0000

Data Index: 1417915 (Python Code)

Max Activating Token Index: 576

Text #9

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3159. Min Act: -0.0000

Data Index: 925677 (C4 (Web Text))

Max Activating Token Index: 502

Text #10

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2879. Min Act: -0.0000

Data Index: 1155171 (C4 (Web Text))

Max Activating Token Index: 775

Text #11

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.3059. Min Act: -0.0000

Data Index: 1294629 (C4 (Web Text))

Max Activating Token Index: 504

Text #12

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2863. Min Act: -0.0000

Data Index: 645466 (C4 (Web Text))

Max Activating Token Index: 548

Text #13

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2807. Min Act: -0.0000

Data Index: 1561434 (Python Code)

Max Activating Token Index: 794

Text #14

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2699. Min Act: -0.0000

Data Index: 1257056 (C4 (Web Text))

Max Activating Token Index: 391

Text #15

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2688. Min Act: -0.0000

Data Index: 1634389 (Python Code)

Max Activating Token Index: 384

Text #16

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2969. Min Act: -0.0000

Data Index: 91135 (C4 (Web Text))

Max Activating Token Index: 478

Text #17

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2865. Min Act: -0.0000

Data Index: 1107209 (C4 (Web Text))

Max Activating Token Index: 452

Text #18

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2889. Min Act: -0.0000

Data Index: 365942 (C4 (Web Text))

Max Activating Token Index: 694

Text #19

Max Range: 0.4120. Min Range: -0.4120

Max Act: 0.2847. Min Act: -0.0000

Data Index: 27324 (C4 (Web Text))

Max Activating Token Index: 115