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

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

Neuron 4766 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.1744. Min Range: -1.1744

Max Act: 1.1744. Min Act: -0.0000

Data Index: 1296596 (C4 (Web Text))

Max Activating Token Index: 47

Text #1

Max Range: 1.1744. Min Range: -1.1744

Max Act: 1.0831. Min Act: -0.0000

Data Index: 10298 (C4 (Web Text))

Max Activating Token Index: 611

Text #2

Max Range: 1.1744. Min Range: -1.1744

Max Act: 1.0290. Min Act: -0.0000

Data Index: 538850 (C4 (Web Text))

Max Activating Token Index: 78

Text #3

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9921. Min Act: -0.0000

Data Index: 999095 (C4 (Web Text))

Max Activating Token Index: 827

Text #4

Max Range: 1.1744. Min Range: -1.1744

Max Act: 1.0132. Min Act: -0.0000

Data Index: 706860 (C4 (Web Text))

Max Activating Token Index: 72

Text #5

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9580. Min Act: -0.0000

Data Index: 907527 (C4 (Web Text))

Max Activating Token Index: 844

Text #6

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9044. Min Act: -0.0001

Data Index: 1579748 (Python Code)

Max Activating Token Index: 626

Text #7

Max Range: 1.1744. Min Range: -1.1744

Max Act: 1.0239. Min Act: -0.0000

Data Index: 315152 (C4 (Web Text))

Max Activating Token Index: 789

Text #8

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9031. Min Act: -0.0000

Data Index: 401018 (C4 (Web Text))

Max Activating Token Index: 72

Text #9

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9554. Min Act: -0.0000

Data Index: 871624 (C4 (Web Text))

Max Activating Token Index: 634

Text #10

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.7204. Min Act: -0.0000

Data Index: 1491213 (Python Code)

Max Activating Token Index: 696

Text #11

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8517. Min Act: -0.0000

Data Index: 1193809 (C4 (Web Text))

Max Activating Token Index: 221

Text #12

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8969. Min Act: -0.0000

Data Index: 79620 (C4 (Web Text))

Max Activating Token Index: 428

Text #13

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8139. Min Act: -0.0000

Data Index: 730284 (C4 (Web Text))

Max Activating Token Index: 623

Text #14

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8432. Min Act: -0.0000

Data Index: 733583 (C4 (Web Text))

Max Activating Token Index: 210

Text #15

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.9048. Min Act: -0.0000

Data Index: 1319478 (C4 (Web Text))

Max Activating Token Index: 667

Text #16

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8842. Min Act: -0.0000

Data Index: 613511 (C4 (Web Text))

Max Activating Token Index: 924

Text #17

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8457. Min Act: -0.0000

Data Index: 1502690 (Python Code)

Max Activating Token Index: 351

Text #18

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.8466. Min Act: -0.0000

Data Index: 596495 (C4 (Web Text))

Max Activating Token Index: 801

Text #19

Max Range: 1.1744. Min Range: -1.1744

Max Act: 0.7112. Min Act: -0.0000

Data Index: 519506 (C4 (Web Text))

Max Activating Token Index: 175