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

Dataset: The Pile

Neuron 450 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 7.6217. Min Range: -7.6217

Max Act: 7.6217. Min Act: -0.0000

Data Index: 988982 (The Pile)

Max Activating Token Index: 1005

Text #1

Max Range: 7.6217. Min Range: -7.6217

Max Act: 7.5117. Min Act: -0.0000

Data Index: 705891 (The Pile)

Max Activating Token Index: 544

Text #2

Max Range: 7.6217. Min Range: -7.6217

Max Act: 7.0696. Min Act: -0.0000

Data Index: 269477 (The Pile)

Max Activating Token Index: 179

Text #3

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.8647. Min Act: -0.0000

Data Index: 162187 (The Pile)

Max Activating Token Index: 974

Text #4

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.7123. Min Act: -0.0000

Data Index: 579834 (The Pile)

Max Activating Token Index: 816

Text #5

Max Range: 7.6217. Min Range: -7.6217

Max Act: 7.3878. Min Act: -0.0000

Data Index: 1744535 (The Pile)

Max Activating Token Index: 700

Text #6

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.7854. Min Act: -0.0000

Data Index: 915221 (The Pile)

Max Activating Token Index: 436

Text #7

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.1373. Min Act: -0.0000

Data Index: 1425684 (The Pile)

Max Activating Token Index: 712

Text #8

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.6555. Min Act: -0.0000

Data Index: 374834 (The Pile)

Max Activating Token Index: 753

Text #9

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.4155. Min Act: -0.0000

Data Index: 1982851 (The Pile)

Max Activating Token Index: 359

Text #10

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.4184. Min Act: -0.0000

Data Index: 1803251 (The Pile)

Max Activating Token Index: 332

Text #11

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.5841. Min Act: -0.0000

Data Index: 978537 (The Pile)

Max Activating Token Index: 881

Text #12

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.4760. Min Act: -0.0000

Data Index: 994663 (The Pile)

Max Activating Token Index: 687

Text #13

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.4796. Min Act: -0.0000

Data Index: 1396239 (The Pile)

Max Activating Token Index: 546

Text #14

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.2533. Min Act: -0.0000

Data Index: 621210 (The Pile)

Max Activating Token Index: 494

Text #15

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.3473. Min Act: -0.0000

Data Index: 1093194 (The Pile)

Max Activating Token Index: 478

Text #16

Max Range: 7.6217. Min Range: -7.6217

Max Act: 5.6883. Min Act: -0.0000

Data Index: 982650 (The Pile)

Max Activating Token Index: 899

Text #17

Max Range: 7.6217. Min Range: -7.6217

Max Act: 6.3785. Min Act: -0.0000

Data Index: 438553 (The Pile)

Max Activating Token Index: 310

Text #18

Max Range: 7.6217. Min Range: -7.6217

Max Act: 5.8956. Min Act: -0.0000

Data Index: 1789520 (The Pile)

Max Activating Token Index: 919

Text #19

Max Range: 7.6217. Min Range: -7.6217

Max Act: 5.7814. Min Act: -0.0000

Data Index: 508856 (The Pile)

Max Activating Token Index: 94