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

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

Neuron 4636 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: 0.6857. Min Range: -0.6857

Max Act: 0.6857. Min Act: -0.0000

Data Index: 828853 (C4 (Web Text))

Max Activating Token Index: 823

Text #1

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.6191. Min Act: -0.0000

Data Index: 1463615 (Python Code)

Max Activating Token Index: 93

Text #2

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5942. Min Act: -0.0000

Data Index: 1488354 (Python Code)

Max Activating Token Index: 418

Text #3

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5714. Min Act: -0.0000

Data Index: 1183013 (C4 (Web Text))

Max Activating Token Index: 319

Text #4

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5970. Min Act: -0.0000

Data Index: 1552236 (Python Code)

Max Activating Token Index: 343

Text #5

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5762. Min Act: -0.0000

Data Index: 522660 (C4 (Web Text))

Max Activating Token Index: 595

Text #6

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5951. Min Act: -0.0000

Data Index: 1617818 (Python Code)

Max Activating Token Index: 907

Text #7

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5733. Min Act: -0.0000

Data Index: 1375189 (Python Code)

Max Activating Token Index: 898

Text #8

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5856. Min Act: -0.0000

Data Index: 591510 (C4 (Web Text))

Max Activating Token Index: 87

Text #9

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5493. Min Act: -0.0000

Data Index: 1069704 (C4 (Web Text))

Max Activating Token Index: 154

Text #10

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5630. Min Act: -0.0000

Data Index: 1037587 (C4 (Web Text))

Max Activating Token Index: 407

Text #11

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5584. Min Act: -0.0000

Data Index: 1491171 (Python Code)

Max Activating Token Index: 638

Text #12

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5578. Min Act: -0.0000

Data Index: 1481262 (Python Code)

Max Activating Token Index: 855

Text #13

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5561. Min Act: -0.0000

Data Index: 1275370 (C4 (Web Text))

Max Activating Token Index: 414

Text #14

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5884. Min Act: -0.0000

Data Index: 618577 (C4 (Web Text))

Max Activating Token Index: 673

Text #15

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5424. Min Act: -0.0000

Data Index: 1385162 (Python Code)

Max Activating Token Index: 959

Text #16

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5438. Min Act: -0.0000

Data Index: 597236 (C4 (Web Text))

Max Activating Token Index: 297

Text #17

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5426. Min Act: -0.0000

Data Index: 952257 (C4 (Web Text))

Max Activating Token Index: 691

Text #18

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5430. Min Act: -0.0000

Data Index: 632817 (C4 (Web Text))

Max Activating Token Index: 188

Text #19

Max Range: 0.6857. Min Range: -0.6857

Max Act: 0.5392. Min Act: -0.0000

Data Index: 1576554 (Python Code)

Max Activating Token Index: 778