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

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

Neuron 635 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: 0.8297. Min Range: -0.8297

Max Act: 0.8297. Min Act: -0.0000

Data Index: 697396 (C4 (Web Text))

Max Activating Token Index: 964

Text #1

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7426. Min Act: -0.0000

Data Index: 640308 (C4 (Web Text))

Max Activating Token Index: 63

Text #2

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7740. Min Act: -0.0000

Data Index: 1498267 (Python Code)

Max Activating Token Index: 38

Text #3

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7862. Min Act: -0.0000

Data Index: 139330 (C4 (Web Text))

Max Activating Token Index: 65

Text #4

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7140. Min Act: -0.0000

Data Index: 1606727 (Python Code)

Max Activating Token Index: 124

Text #5

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7613. Min Act: -0.0000

Data Index: 1121170 (C4 (Web Text))

Max Activating Token Index: 656

Text #6

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6665. Min Act: -0.0000

Data Index: 1488449 (Python Code)

Max Activating Token Index: 68

Text #7

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6464. Min Act: -0.0000

Data Index: 1390529 (Python Code)

Max Activating Token Index: 837

Text #8

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7103. Min Act: -0.0000

Data Index: 1204855 (C4 (Web Text))

Max Activating Token Index: 455

Text #9

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6922. Min Act: -0.0000

Data Index: 1094497 (C4 (Web Text))

Max Activating Token Index: 635

Text #10

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6808. Min Act: -0.0000

Data Index: 1121902 (C4 (Web Text))

Max Activating Token Index: 909

Text #11

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.5729. Min Act: -0.0000

Data Index: 1250042 (C4 (Web Text))

Max Activating Token Index: 591

Text #12

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.7019. Min Act: -0.0000

Data Index: 172436 (C4 (Web Text))

Max Activating Token Index: 348

Text #13

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6036. Min Act: -0.0000

Data Index: 363757 (C4 (Web Text))

Max Activating Token Index: 644

Text #14

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6594. Min Act: -0.0000

Data Index: 1080599 (C4 (Web Text))

Max Activating Token Index: 840

Text #15

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6674. Min Act: -0.0000

Data Index: 10404 (C4 (Web Text))

Max Activating Token Index: 805

Text #16

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6618. Min Act: -0.0000

Data Index: 996712 (C4 (Web Text))

Max Activating Token Index: 98

Text #17

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6588. Min Act: -0.0000

Data Index: 1631049 (Python Code)

Max Activating Token Index: 859

Text #18

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6394. Min Act: -0.0000

Data Index: 1003763 (C4 (Web Text))

Max Activating Token Index: 375

Text #19

Max Range: 0.8297. Min Range: -0.8297

Max Act: 0.6682. Min Act: -0.0000

Data Index: 278535 (C4 (Web Text))

Max Activating Token Index: 395