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

Dataset: The Pile

Neuron 21 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 7.0661. Min Range: -7.0661

Max Act: 7.0661. Min Act: -0.0000

Data Index: 704529 (The Pile)

Max Activating Token Index: 953

Text #1

Max Range: 7.0661. Min Range: -7.0661

Max Act: 6.6817. Min Act: -0.0000

Data Index: 1532185 (The Pile)

Max Activating Token Index: 978

Text #2

Max Range: 7.0661. Min Range: -7.0661

Max Act: 6.6371. Min Act: -0.0000

Data Index: 1690823 (The Pile)

Max Activating Token Index: 705

Text #3

Max Range: 7.0661. Min Range: -7.0661

Max Act: 6.0604. Min Act: -0.0000

Data Index: 585021 (The Pile)

Max Activating Token Index: 28

Text #4

Max Range: 7.0661. Min Range: -7.0661

Max Act: 6.0147. Min Act: -0.0000

Data Index: 593905 (The Pile)

Max Activating Token Index: 609

Text #5

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.9491. Min Act: -0.0000

Data Index: 347903 (The Pile)

Max Activating Token Index: 111

Text #6

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.6571. Min Act: -0.0000

Data Index: 1865062 (The Pile)

Max Activating Token Index: 454

Text #7

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.5868. Min Act: -0.0000

Data Index: 282663 (The Pile)

Max Activating Token Index: 704

Text #8

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.6219. Min Act: -0.0000

Data Index: 1797691 (The Pile)

Max Activating Token Index: 377

Text #9

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.5431. Min Act: -0.0000

Data Index: 1266550 (The Pile)

Max Activating Token Index: 251

Text #10

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.6731. Min Act: -0.0000

Data Index: 87109 (The Pile)

Max Activating Token Index: 36

Text #11

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.4422. Min Act: -0.0000

Data Index: 354968 (The Pile)

Max Activating Token Index: 437

Text #12

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.3986. Min Act: -0.0000

Data Index: 1797329 (The Pile)

Max Activating Token Index: 304

Text #13

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.5814. Min Act: -0.0000

Data Index: 239937 (The Pile)

Max Activating Token Index: 1016

Text #14

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.1480. Min Act: -0.0000

Data Index: 1384905 (The Pile)

Max Activating Token Index: 1023

Text #15

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.0551. Min Act: -0.0000

Data Index: 335859 (The Pile)

Max Activating Token Index: 462

Text #16

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.2695. Min Act: -0.0000

Data Index: 918102 (The Pile)

Max Activating Token Index: 675

Text #17

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.0330. Min Act: -0.0000

Data Index: 230215 (The Pile)

Max Activating Token Index: 669

Text #18

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.2121. Min Act: -0.0000

Data Index: 373649 (The Pile)

Max Activating Token Index: 634

Text #19

Max Range: 7.0661. Min Range: -7.0661

Max Act: 5.1895. Min Act: -0.0000

Data Index: 918144 (The Pile)

Max Activating Token Index: 641