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

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

Neuron 17 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.5866. Min Act: -0.0000

Data Index: 71840 (C4 (Web Text))

Max Activating Token Index: 962

Text #1

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.1433. Min Act: -0.0000

Data Index: 1341008 (C4 (Web Text))

Max Activating Token Index: 279

Text #2

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.2984. Min Act: -0.0000

Data Index: 836949 (C4 (Web Text))

Max Activating Token Index: 187

Text #3

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.0963. Min Act: -0.0000

Data Index: 1093340 (C4 (Web Text))

Max Activating Token Index: 74

Text #4

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.0332. Min Act: -0.0000

Data Index: 285080 (C4 (Web Text))

Max Activating Token Index: 283

Text #5

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.0436. Min Act: -0.0000

Data Index: 222659 (C4 (Web Text))

Max Activating Token Index: 692

Text #6

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.0398. Min Act: -0.0000

Data Index: 700171 (C4 (Web Text))

Max Activating Token Index: 156

Text #7

Max Range: 2.5866. Min Range: -2.5866

Max Act: 2.0759. Min Act: -0.0000

Data Index: 631435 (C4 (Web Text))

Max Activating Token Index: 370

Text #8

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.9538. Min Act: 0.0003

Data Index: 1093277 (C4 (Web Text))

Max Activating Token Index: 495

Text #9

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8685. Min Act: -0.0000

Data Index: 345021 (C4 (Web Text))

Max Activating Token Index: 338

Text #10

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.7667. Min Act: -0.0000

Data Index: 285986 (C4 (Web Text))

Max Activating Token Index: 689

Text #11

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8485. Min Act: -0.0000

Data Index: 638333 (C4 (Web Text))

Max Activating Token Index: 727

Text #12

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8399. Min Act: -0.0000

Data Index: 657976 (C4 (Web Text))

Max Activating Token Index: 210

Text #13

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8420. Min Act: 0.0003

Data Index: 538109 (C4 (Web Text))

Max Activating Token Index: 582

Text #14

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.7107. Min Act: -0.0000

Data Index: 870213 (C4 (Web Text))

Max Activating Token Index: 757

Text #15

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8044. Min Act: -0.0000

Data Index: 813920 (C4 (Web Text))

Max Activating Token Index: 665

Text #16

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.8012. Min Act: -0.0000

Data Index: 903590 (C4 (Web Text))

Max Activating Token Index: 408

Text #17

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.7416. Min Act: -0.0000

Data Index: 1616520 (Python Code)

Max Activating Token Index: 445

Text #18

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.7411. Min Act: 0.0000

Data Index: 670022 (C4 (Web Text))

Max Activating Token Index: 193

Text #19

Max Range: 2.5866. Min Range: -2.5866

Max Act: 1.6284. Min Act: 0.0000

Data Index: 1020625 (C4 (Web Text))

Max Activating Token Index: 232