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

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

Neuron 1589 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: 0.0744. Min Range: -0.0744

Max Act: 0.0744. Min Act: -0.0002

Data Index: 1429368 (Python Code)

Max Activating Token Index: 955

Text #1

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0670. Min Act: -0.0001

Data Index: 120634 (C4 (Web Text))

Max Activating Token Index: 220

Text #2

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0655. Min Act: -0.0001

Data Index: 444770 (C4 (Web Text))

Max Activating Token Index: 513

Text #3

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0618. Min Act: -0.0001

Data Index: 228916 (C4 (Web Text))

Max Activating Token Index: 875

Text #4

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0617. Min Act: -0.0001

Data Index: 190623 (C4 (Web Text))

Max Activating Token Index: 483

Text #5

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0573. Min Act: -0.0001

Data Index: 1051128 (C4 (Web Text))

Max Activating Token Index: 368

Text #6

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0568. Min Act: -0.0002

Data Index: 1388850 (Python Code)

Max Activating Token Index: 965

Text #7

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0610. Min Act: -0.0000

Data Index: 1074440 (C4 (Web Text))

Max Activating Token Index: 995

Text #8

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0574. Min Act: -0.0000

Data Index: 602668 (C4 (Web Text))

Max Activating Token Index: 628

Text #9

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0579. Min Act: -0.0001

Data Index: 221418 (C4 (Web Text))

Max Activating Token Index: 941

Text #10

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0552. Min Act: -0.0000

Data Index: 422948 (C4 (Web Text))

Max Activating Token Index: 587

Text #11

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0565. Min Act: -0.0001

Data Index: 566179 (C4 (Web Text))

Max Activating Token Index: 727

Text #12

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0533. Min Act: -0.0002

Data Index: 290300 (C4 (Web Text))

Max Activating Token Index: 121

Text #13

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0527. Min Act: -0.0001

Data Index: 1505272 (Python Code)

Max Activating Token Index: 613

Text #14

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0510. Min Act: -0.0001

Data Index: 1409280 (Python Code)

Max Activating Token Index: 677

Text #15

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0546. Min Act: -0.0001

Data Index: 849387 (C4 (Web Text))

Max Activating Token Index: 930

Text #16

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0556. Min Act: -0.0001

Data Index: 1382302 (Python Code)

Max Activating Token Index: 643

Text #17

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0537. Min Act: -0.0000

Data Index: 285469 (C4 (Web Text))

Max Activating Token Index: 532

Text #18

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0545. Min Act: -0.0001

Data Index: 1494943 (Python Code)

Max Activating Token Index: 395

Text #19

Max Range: 0.0744. Min Range: -0.0744

Max Act: 0.0542. Min Act: -0.0001

Data Index: 795006 (C4 (Web Text))

Max Activating Token Index: 718