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

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

Neuron 1584 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0589. Min Act: -0.0003

Data Index: 1554512 (Python Code)

Max Activating Token Index: 626

Text #1

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0524. Min Act: -0.0003

Data Index: 621909 (C4 (Web Text))

Max Activating Token Index: 896

Text #2

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0546. Min Act: -0.0003

Data Index: 1104270 (C4 (Web Text))

Max Activating Token Index: 600

Text #3

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0540. Min Act: -0.0003

Data Index: 829931 (C4 (Web Text))

Max Activating Token Index: 848

Text #4

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0534. Min Act: -0.0003

Data Index: 1415858 (Python Code)

Max Activating Token Index: 755

Text #5

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0524. Min Act: -0.0003

Data Index: 1372078 (Python Code)

Max Activating Token Index: 629

Text #6

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0515. Min Act: -0.0003

Data Index: 621082 (C4 (Web Text))

Max Activating Token Index: 869

Text #7

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0509. Min Act: -0.0003

Data Index: 1514174 (Python Code)

Max Activating Token Index: 602

Text #8

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0482. Min Act: -0.0003

Data Index: 1127837 (C4 (Web Text))

Max Activating Token Index: 651

Text #9

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0502. Min Act: -0.0003

Data Index: 850458 (C4 (Web Text))

Max Activating Token Index: 871

Text #10

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0502. Min Act: -0.0003

Data Index: 269410 (C4 (Web Text))

Max Activating Token Index: 158

Text #11

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0509. Min Act: -0.0003

Data Index: 1453542 (Python Code)

Max Activating Token Index: 923

Text #12

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0514. Min Act: -0.0003

Data Index: 429373 (C4 (Web Text))

Max Activating Token Index: 253

Text #13

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0503. Min Act: -0.0003

Data Index: 623928 (C4 (Web Text))

Max Activating Token Index: 926

Text #14

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0498. Min Act: -0.0003

Data Index: 18117 (C4 (Web Text))

Max Activating Token Index: 801

Text #15

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0481. Min Act: -0.0003

Data Index: 209430 (C4 (Web Text))

Max Activating Token Index: 36

Text #16

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0488. Min Act: -0.0003

Data Index: 1510084 (Python Code)

Max Activating Token Index: 988

Text #17

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0490. Min Act: -0.0003

Data Index: 1052023 (C4 (Web Text))

Max Activating Token Index: 144

Text #18

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0491. Min Act: -0.0003

Data Index: 1249357 (C4 (Web Text))

Max Activating Token Index: 90

Text #19

Max Range: 0.0589. Min Range: -0.0589

Max Act: 0.0481. Min Act: -0.0003

Data Index: 405513 (C4 (Web Text))

Max Activating Token Index: 807