Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3999 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.1384. Min Act: -0.0000

Data Index: 58033 (C4 (Web Text))

Max Activating Token Index: 5

Text #1

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.1131. Min Act: -0.0001

Data Index: 177136 (C4 (Web Text))

Max Activating Token Index: 402

Text #2

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.1013. Min Act: -0.0001

Data Index: 785914 (C4 (Web Text))

Max Activating Token Index: 256

Text #3

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0922. Min Act: -0.0001

Data Index: 1092645 (C4 (Web Text))

Max Activating Token Index: 570

Text #4

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0931. Min Act: -0.0001

Data Index: 867483 (C4 (Web Text))

Max Activating Token Index: 37

Text #5

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0992. Min Act: -0.0000

Data Index: 329360 (C4 (Web Text))

Max Activating Token Index: 149

Text #6

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0888. Min Act: -0.0001

Data Index: 759635 (C4 (Web Text))

Max Activating Token Index: 141

Text #7

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0835. Min Act: -0.0001

Data Index: 1438838 (Python Code)

Max Activating Token Index: 12

Text #8

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0848. Min Act: -0.0000

Data Index: 507964 (C4 (Web Text))

Max Activating Token Index: 337

Text #9

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0824. Min Act: -0.0001

Data Index: 5122 (C4 (Web Text))

Max Activating Token Index: 236

Text #10

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0822. Min Act: -0.0001

Data Index: 885291 (C4 (Web Text))

Max Activating Token Index: 12

Text #11

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0780. Min Act: -0.0001

Data Index: 259074 (C4 (Web Text))

Max Activating Token Index: 4

Text #12

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0746. Min Act: -0.0001

Data Index: 1452805 (Python Code)

Max Activating Token Index: 614

Text #13

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0746. Min Act: -0.0001

Data Index: 7441 (C4 (Web Text))

Max Activating Token Index: 73

Text #14

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0738. Min Act: -0.0001

Data Index: 1138865 (C4 (Web Text))

Max Activating Token Index: 721

Text #15

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0706. Min Act: -0.0001

Data Index: 1167756 (C4 (Web Text))

Max Activating Token Index: 452

Text #16

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0764. Min Act: -0.0001

Data Index: 1013987 (C4 (Web Text))

Max Activating Token Index: 129

Text #17

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0726. Min Act: -0.0001

Data Index: 1284721 (C4 (Web Text))

Max Activating Token Index: 605

Text #18

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0688. Min Act: -0.0000

Data Index: 812135 (C4 (Web Text))

Max Activating Token Index: 62

Text #19

Max Range: 0.1384. Min Range: -0.1384

Max Act: 0.0710. Min Act: -0.0001

Data Index: 662652 (C4 (Web Text))

Max Activating Token Index: 387