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

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

Neuron 3199 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.7970. Min Act: -0.0000

Data Index: 14110 (C4 (Web Text))

Max Activating Token Index: 84

Text #1

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.3480. Min Act: -0.0000

Data Index: 1457385 (Python Code)

Max Activating Token Index: 886

Text #2

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.4819. Min Act: -0.0000

Data Index: 935103 (C4 (Web Text))

Max Activating Token Index: 566

Text #3

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.2897. Min Act: -0.0000

Data Index: 1386379 (Python Code)

Max Activating Token Index: 1010

Text #4

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1010. Min Act: -0.0000

Data Index: 405152 (C4 (Web Text))

Max Activating Token Index: 546

Text #5

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.2331. Min Act: -0.0000

Data Index: 30376 (C4 (Web Text))

Max Activating Token Index: 652

Text #6

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.3368. Min Act: -0.0000

Data Index: 1069321 (C4 (Web Text))

Max Activating Token Index: 753

Text #7

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1127. Min Act: -0.0000

Data Index: 1220092 (C4 (Web Text))

Max Activating Token Index: 637

Text #8

Max Range: 2.7970. Min Range: -2.7970

Max Act: 1.5944. Min Act: -0.0000

Data Index: 1312773 (C4 (Web Text))

Max Activating Token Index: 380

Text #9

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0746. Min Act: -0.0000

Data Index: 1192311 (C4 (Web Text))

Max Activating Token Index: 343

Text #10

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0856. Min Act: -0.0000

Data Index: 1288788 (C4 (Web Text))

Max Activating Token Index: 995

Text #11

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1680. Min Act: -0.0000

Data Index: 1308775 (C4 (Web Text))

Max Activating Token Index: 230

Text #12

Max Range: 2.7970. Min Range: -2.7970

Max Act: 1.9416. Min Act: -0.0000

Data Index: 1353895 (C4 (Web Text))

Max Activating Token Index: 367

Text #13

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0500. Min Act: -0.0000

Data Index: 1446301 (Python Code)

Max Activating Token Index: 200

Text #14

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0556. Min Act: -0.0000

Data Index: 1178709 (C4 (Web Text))

Max Activating Token Index: 656

Text #15

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0476. Min Act: -0.0000

Data Index: 35590 (C4 (Web Text))

Max Activating Token Index: 852

Text #16

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1409. Min Act: -0.0000

Data Index: 702995 (C4 (Web Text))

Max Activating Token Index: 862

Text #17

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.0285. Min Act: -0.0000

Data Index: 644719 (C4 (Web Text))

Max Activating Token Index: 582

Text #18

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1312. Min Act: -0.0000

Data Index: 786389 (C4 (Web Text))

Max Activating Token Index: 766

Text #19

Max Range: 2.7970. Min Range: -2.7970

Max Act: 2.1989. Min Act: -0.0000

Data Index: 380586 (C4 (Web Text))

Max Activating Token Index: 916