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

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

Neuron 3605 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6796. Min Act: -0.0000

Data Index: 928972 (C4 (Web Text))

Max Activating Token Index: 449

Text #1

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6918. Min Act: -0.0000

Data Index: 51184 (C4 (Web Text))

Max Activating Token Index: 902

Text #2

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6837. Min Act: -0.0000

Data Index: 1296026 (C4 (Web Text))

Max Activating Token Index: 391

Text #3

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5952. Min Act: -0.0000

Data Index: 307151 (C4 (Web Text))

Max Activating Token Index: 65

Text #4

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6189. Min Act: -0.0000

Data Index: 938839 (C4 (Web Text))

Max Activating Token Index: 221

Text #5

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6189. Min Act: -0.0000

Data Index: 608127 (C4 (Web Text))

Max Activating Token Index: 124

Text #6

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5956. Min Act: -0.0000

Data Index: 693642 (C4 (Web Text))

Max Activating Token Index: 493

Text #7

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6681. Min Act: -0.0000

Data Index: 758255 (C4 (Web Text))

Max Activating Token Index: 258

Text #8

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5353. Min Act: -0.0000

Data Index: 477176 (C4 (Web Text))

Max Activating Token Index: 693

Text #9

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5590. Min Act: -0.0000

Data Index: 266818 (C4 (Web Text))

Max Activating Token Index: 234

Text #10

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5700. Min Act: -0.0000

Data Index: 1152451 (C4 (Web Text))

Max Activating Token Index: 126

Text #11

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5614. Min Act: -0.0000

Data Index: 560889 (C4 (Web Text))

Max Activating Token Index: 427

Text #12

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.4985. Min Act: -0.0000

Data Index: 64631 (C4 (Web Text))

Max Activating Token Index: 966

Text #13

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.6098. Min Act: -0.0000

Data Index: 388704 (C4 (Web Text))

Max Activating Token Index: 631

Text #14

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5089. Min Act: -0.0000

Data Index: 600137 (C4 (Web Text))

Max Activating Token Index: 857

Text #15

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.4608. Min Act: -0.0000

Data Index: 495630 (C4 (Web Text))

Max Activating Token Index: 430

Text #16

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.4414. Min Act: -0.0000

Data Index: 1434021 (Python Code)

Max Activating Token Index: 900

Text #17

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.5036. Min Act: -0.0000

Data Index: 1641856 (Python Code)

Max Activating Token Index: 840

Text #18

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.4655. Min Act: -0.0000

Data Index: 675905 (C4 (Web Text))

Max Activating Token Index: 967

Text #19

Max Range: 0.6918. Min Range: -0.6918

Max Act: 0.4583. Min Act: -0.0000

Data Index: 1368796 (Python Code)

Max Activating Token Index: 112