Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1306 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0768. Min Act: -0.0002

Data Index: 773751 (C4 (Web Text))

Max Activating Token Index: 993

Text #1

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0761. Min Act: -0.0002

Data Index: 676638 (C4 (Web Text))

Max Activating Token Index: 572

Text #2

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0686. Min Act: -0.0002

Data Index: 179496 (C4 (Web Text))

Max Activating Token Index: 957

Text #3

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0700. Min Act: -0.0002

Data Index: 114382 (C4 (Web Text))

Max Activating Token Index: 792

Text #4

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0653. Min Act: -0.0002

Data Index: 987474 (C4 (Web Text))

Max Activating Token Index: 13

Text #5

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0642. Min Act: -0.0001

Data Index: 281382 (C4 (Web Text))

Max Activating Token Index: 648

Text #6

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0628. Min Act: -0.0002

Data Index: 946132 (C4 (Web Text))

Max Activating Token Index: 351

Text #7

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0660. Min Act: -0.0002

Data Index: 892474 (C4 (Web Text))

Max Activating Token Index: 118

Text #8

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0644. Min Act: -0.0002

Data Index: 40638 (C4 (Web Text))

Max Activating Token Index: 307

Text #9

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0627. Min Act: -0.0001

Data Index: 925859 (C4 (Web Text))

Max Activating Token Index: 994

Text #10

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0619. Min Act: -0.0002

Data Index: 840603 (C4 (Web Text))

Max Activating Token Index: 321

Text #11

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0615. Min Act: -0.0002

Data Index: 55120 (C4 (Web Text))

Max Activating Token Index: 752

Text #12

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0615. Min Act: -0.0002

Data Index: 200560 (C4 (Web Text))

Max Activating Token Index: 683

Text #13

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0615. Min Act: -0.0001

Data Index: 73677 (C4 (Web Text))

Max Activating Token Index: 712

Text #14

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0603. Min Act: -0.0002

Data Index: 939702 (C4 (Web Text))

Max Activating Token Index: 403

Text #15

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0618. Min Act: -0.0002

Data Index: 968660 (C4 (Web Text))

Max Activating Token Index: 274

Text #16

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0600. Min Act: -0.0002

Data Index: 822360 (C4 (Web Text))

Max Activating Token Index: 817

Text #17

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0608. Min Act: -0.0002

Data Index: 107846 (C4 (Web Text))

Max Activating Token Index: 494

Text #18

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0593. Min Act: -0.0002

Data Index: 466526 (C4 (Web Text))

Max Activating Token Index: 593

Text #19

Max Range: 0.0768. Min Range: -0.0768

Max Act: 0.0587. Min Act: -0.0002

Data Index: 1074261 (C4 (Web Text))

Max Activating Token Index: 207