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

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

Neuron 2008 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.1002. Min Range: -1.1002

Max Act: 1.0403. Min Act: -0.0000

Data Index: 1092567 (C4 (Web Text))

Max Activating Token Index: 970

Text #1

Max Range: 1.1002. Min Range: -1.1002

Max Act: 1.1002. Min Act: -0.0000

Data Index: 141371 (C4 (Web Text))

Max Activating Token Index: 177

Text #2

Max Range: 1.1002. Min Range: -1.1002

Max Act: 1.0353. Min Act: -0.0000

Data Index: 1241638 (C4 (Web Text))

Max Activating Token Index: 443

Text #3

Max Range: 1.1002. Min Range: -1.1002

Max Act: 1.0255. Min Act: -0.0000

Data Index: 566416 (C4 (Web Text))

Max Activating Token Index: 667

Text #4

Max Range: 1.1002. Min Range: -1.1002

Max Act: 1.0049. Min Act: -0.0000

Data Index: 739727 (C4 (Web Text))

Max Activating Token Index: 396

Text #5

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.9789. Min Act: -0.0000

Data Index: 139654 (C4 (Web Text))

Max Activating Token Index: 127

Text #6

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8527. Min Act: -0.0000

Data Index: 1120536 (C4 (Web Text))

Max Activating Token Index: 764

Text #7

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.9072. Min Act: -0.0000

Data Index: 164078 (C4 (Web Text))

Max Activating Token Index: 488

Text #8

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.9004. Min Act: -0.0000

Data Index: 330922 (C4 (Web Text))

Max Activating Token Index: 440

Text #9

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.9027. Min Act: -0.0000

Data Index: 1566840 (Python Code)

Max Activating Token Index: 74

Text #10

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8958. Min Act: -0.0000

Data Index: 803904 (C4 (Web Text))

Max Activating Token Index: 1004

Text #11

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8886. Min Act: -0.0000

Data Index: 556935 (C4 (Web Text))

Max Activating Token Index: 124

Text #12

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8555. Min Act: -0.0000

Data Index: 678130 (C4 (Web Text))

Max Activating Token Index: 78

Text #13

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8785. Min Act: -0.0000

Data Index: 650517 (C4 (Web Text))

Max Activating Token Index: 429

Text #14

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8718. Min Act: -0.0000

Data Index: 150213 (C4 (Web Text))

Max Activating Token Index: 563

Text #15

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.9019. Min Act: -0.0000

Data Index: 1246671 (C4 (Web Text))

Max Activating Token Index: 346

Text #16

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8653. Min Act: -0.0000

Data Index: 1481685 (Python Code)

Max Activating Token Index: 287

Text #17

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8782. Min Act: -0.0000

Data Index: 383664 (C4 (Web Text))

Max Activating Token Index: 446

Text #18

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8248. Min Act: -0.0000

Data Index: 1284950 (C4 (Web Text))

Max Activating Token Index: 814

Text #19

Max Range: 1.1002. Min Range: -1.1002

Max Act: 0.8466. Min Act: -0.0000

Data Index: 436867 (C4 (Web Text))

Max Activating Token Index: 28