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

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

Neuron 1035 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0588. Min Act: -0.0005

Data Index: 259735 (C4 (Web Text))

Max Activating Token Index: 465

Text #1

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0575. Min Act: -0.0005

Data Index: 1159366 (C4 (Web Text))

Max Activating Token Index: 108

Text #2

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0548. Min Act: -0.0005

Data Index: 417945 (C4 (Web Text))

Max Activating Token Index: 115

Text #3

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0562. Min Act: -0.0005

Data Index: 1423309 (Python Code)

Max Activating Token Index: 198

Text #4

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0473. Min Act: -0.0005

Data Index: 503671 (C4 (Web Text))

Max Activating Token Index: 1003

Text #5

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0547. Min Act: -0.0005

Data Index: 958751 (C4 (Web Text))

Max Activating Token Index: 952

Text #6

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0545. Min Act: -0.0005

Data Index: 878736 (C4 (Web Text))

Max Activating Token Index: 607

Text #7

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0550. Min Act: -0.0005

Data Index: 808090 (C4 (Web Text))

Max Activating Token Index: 485

Text #8

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0535. Min Act: -0.0005

Data Index: 585531 (C4 (Web Text))

Max Activating Token Index: 303

Text #9

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0572. Min Act: -0.0005

Data Index: 1139900 (C4 (Web Text))

Max Activating Token Index: 870

Text #10

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0511. Min Act: -0.0005

Data Index: 1309401 (C4 (Web Text))

Max Activating Token Index: 190

Text #11

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0529. Min Act: -0.0005

Data Index: 331014 (C4 (Web Text))

Max Activating Token Index: 165

Text #12

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0517. Min Act: -0.0005

Data Index: 877760 (C4 (Web Text))

Max Activating Token Index: 538

Text #13

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0496. Min Act: -0.0005

Data Index: 949275 (C4 (Web Text))

Max Activating Token Index: 417

Text #14

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0503. Min Act: -0.0005

Data Index: 679744 (C4 (Web Text))

Max Activating Token Index: 757

Text #15

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0509. Min Act: -0.0005

Data Index: 1262038 (C4 (Web Text))

Max Activating Token Index: 534

Text #16

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0497. Min Act: -0.0005

Data Index: 443217 (C4 (Web Text))

Max Activating Token Index: 41

Text #17

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0470. Min Act: -0.0005

Data Index: 1619717 (Python Code)

Max Activating Token Index: 298

Text #18

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0503. Min Act: -0.0005

Data Index: 1540648 (Python Code)

Max Activating Token Index: 663

Text #19

Max Range: 0.0588. Min Range: -0.0588

Max Act: 0.0486. Min Act: -0.0005

Data Index: 640430 (C4 (Web Text))

Max Activating Token Index: 536