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

Dataset: The Pile

Neuron 3578 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 8.1618. Min Range: -8.1618

Max Act: 8.1618. Min Act: -0.0000

Data Index: 1910637 (The Pile)

Max Activating Token Index: 599

Text #1

Max Range: 8.1618. Min Range: -8.1618

Max Act: 7.0230. Min Act: -0.0000

Data Index: 1334382 (The Pile)

Max Activating Token Index: 433

Text #2

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.9370. Min Act: -0.0000

Data Index: 734039 (The Pile)

Max Activating Token Index: 210

Text #3

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.6880. Min Act: -0.0000

Data Index: 725905 (The Pile)

Max Activating Token Index: 483

Text #4

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.5007. Min Act: -0.0000

Data Index: 924549 (The Pile)

Max Activating Token Index: 289

Text #5

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.6021. Min Act: -0.0000

Data Index: 34574 (The Pile)

Max Activating Token Index: 881

Text #6

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.7898. Min Act: -0.0000

Data Index: 1148516 (The Pile)

Max Activating Token Index: 405

Text #7

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.6736. Min Act: -0.0000

Data Index: 195945 (The Pile)

Max Activating Token Index: 876

Text #8

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.6442. Min Act: -0.0000

Data Index: 23049 (The Pile)

Max Activating Token Index: 608

Text #9

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.7975. Min Act: -0.0000

Data Index: 788618 (The Pile)

Max Activating Token Index: 427

Text #10

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.6218. Min Act: -0.0000

Data Index: 771312 (The Pile)

Max Activating Token Index: 144

Text #11

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.2921. Min Act: -0.0000

Data Index: 201220 (The Pile)

Max Activating Token Index: 808

Text #12

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.2640. Min Act: -0.0000

Data Index: 1352558 (The Pile)

Max Activating Token Index: 156

Text #13

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.2183. Min Act: -0.0000

Data Index: 354765 (The Pile)

Max Activating Token Index: 463

Text #14

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.1886. Min Act: -0.0000

Data Index: 694175 (The Pile)

Max Activating Token Index: 866

Text #15

Max Range: 8.1618. Min Range: -8.1618

Max Act: 5.7282. Min Act: -0.0000

Data Index: 191631 (The Pile)

Max Activating Token Index: 915

Text #16

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.1291. Min Act: -0.0000

Data Index: 1940720 (The Pile)

Max Activating Token Index: 812

Text #17

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.0309. Min Act: -0.0000

Data Index: 846310 (The Pile)

Max Activating Token Index: 156

Text #18

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.1778. Min Act: -0.0000

Data Index: 1815541 (The Pile)

Max Activating Token Index: 1004

Text #19

Max Range: 8.1618. Min Range: -8.1618

Max Act: 6.3735. Min Act: -0.0000

Data Index: 1440590 (The Pile)

Max Activating Token Index: 665