Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1792 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0664. Min Act: -0.0002

Data Index: 1002349 (C4 (Web Text))

Max Activating Token Index: 101

Text #1

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0666. Min Act: -0.0002

Data Index: 1192466 (C4 (Web Text))

Max Activating Token Index: 103

Text #2

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0614. Min Act: -0.0001

Data Index: 945140 (C4 (Web Text))

Max Activating Token Index: 667

Text #3

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0615. Min Act: -0.0002

Data Index: 1033744 (C4 (Web Text))

Max Activating Token Index: 143

Text #4

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0608. Min Act: -0.0001

Data Index: 86638 (C4 (Web Text))

Max Activating Token Index: 37

Text #5

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0612. Min Act: -0.0002

Data Index: 1165633 (C4 (Web Text))

Max Activating Token Index: 589

Text #6

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0588. Min Act: -0.0001

Data Index: 294945 (C4 (Web Text))

Max Activating Token Index: 215

Text #7

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0602. Min Act: -0.0002

Data Index: 1140085 (C4 (Web Text))

Max Activating Token Index: 934

Text #8

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0600. Min Act: -0.0002

Data Index: 932940 (C4 (Web Text))

Max Activating Token Index: 432

Text #9

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0585. Min Act: -0.0002

Data Index: 911836 (C4 (Web Text))

Max Activating Token Index: 995

Text #10

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0589. Min Act: -0.0002

Data Index: 25325 (C4 (Web Text))

Max Activating Token Index: 262

Text #11

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0590. Min Act: -0.0002

Data Index: 733668 (C4 (Web Text))

Max Activating Token Index: 20

Text #12

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0586. Min Act: -0.0002

Data Index: 283814 (C4 (Web Text))

Max Activating Token Index: 77

Text #13

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0579. Min Act: -0.0001

Data Index: 915124 (C4 (Web Text))

Max Activating Token Index: 824

Text #14

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0560. Min Act: -0.0002

Data Index: 1153212 (C4 (Web Text))

Max Activating Token Index: 448

Text #15

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0575. Min Act: -0.0002

Data Index: 709789 (C4 (Web Text))

Max Activating Token Index: 55

Text #16

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0585. Min Act: -0.0002

Data Index: 1131593 (C4 (Web Text))

Max Activating Token Index: 1007

Text #17

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0569. Min Act: -0.0002

Data Index: 1544576 (Python Code)

Max Activating Token Index: 186

Text #18

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0577. Min Act: -0.0002

Data Index: 182357 (C4 (Web Text))

Max Activating Token Index: 990

Text #19

Max Range: 0.0666. Min Range: -0.0666

Max Act: 0.0564. Min Act: -0.0002

Data Index: 882381 (C4 (Web Text))

Max Activating Token Index: 434