Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 388 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0635. Min Act: -0.0006

Data Index: 799919 (C4 (Web Text))

Max Activating Token Index: 539

Text #1

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0622. Min Act: -0.0006

Data Index: 705362 (C4 (Web Text))

Max Activating Token Index: 506

Text #2

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0622. Min Act: -0.0006

Data Index: 261787 (C4 (Web Text))

Max Activating Token Index: 617

Text #3

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0620. Min Act: -0.0006

Data Index: 821497 (C4 (Web Text))

Max Activating Token Index: 593

Text #4

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0589. Min Act: -0.0006

Data Index: 54876 (C4 (Web Text))

Max Activating Token Index: 654

Text #5

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0618. Min Act: -0.0006

Data Index: 1031193 (C4 (Web Text))

Max Activating Token Index: 244

Text #6

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0584. Min Act: -0.0006

Data Index: 972941 (C4 (Web Text))

Max Activating Token Index: 800

Text #7

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0581. Min Act: -0.0006

Data Index: 177333 (C4 (Web Text))

Max Activating Token Index: 39

Text #8

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0553. Min Act: -0.0006

Data Index: 897568 (C4 (Web Text))

Max Activating Token Index: 728

Text #9

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0575. Min Act: -0.0006

Data Index: 1287235 (C4 (Web Text))

Max Activating Token Index: 394

Text #10

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0573. Min Act: -0.0006

Data Index: 536106 (C4 (Web Text))

Max Activating Token Index: 765

Text #11

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0541. Min Act: -0.0006

Data Index: 36068 (C4 (Web Text))

Max Activating Token Index: 378

Text #12

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0552. Min Act: -0.0006

Data Index: 1168320 (C4 (Web Text))

Max Activating Token Index: 762

Text #13

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0571. Min Act: -0.0006

Data Index: 875578 (C4 (Web Text))

Max Activating Token Index: 689

Text #14

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0545. Min Act: -0.0006

Data Index: 1043212 (C4 (Web Text))

Max Activating Token Index: 288

Text #15

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0557. Min Act: -0.0006

Data Index: 1110334 (C4 (Web Text))

Max Activating Token Index: 31

Text #16

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0568. Min Act: -0.0006

Data Index: 456040 (C4 (Web Text))

Max Activating Token Index: 332

Text #17

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0552. Min Act: -0.0006

Data Index: 23650 (C4 (Web Text))

Max Activating Token Index: 470

Text #18

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0541. Min Act: -0.0006

Data Index: 1356468 (C4 (Web Text))

Max Activating Token Index: 588

Text #19

Max Range: 0.0635. Min Range: -0.0635

Max Act: 0.0536. Min Act: -0.0006

Data Index: 412713 (C4 (Web Text))

Max Activating Token Index: 902