Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2679 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0515. Min Act: -0.0001

Data Index: 344924 (C4 (Web Text))

Max Activating Token Index: 588

Text #1

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0484. Min Act: -0.0001

Data Index: 698712 (C4 (Web Text))

Max Activating Token Index: 968

Text #2

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0512. Min Act: -0.0001

Data Index: 1476170 (Python Code)

Max Activating Token Index: 850

Text #3

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0445. Min Act: -0.0001

Data Index: 1326519 (C4 (Web Text))

Max Activating Token Index: 427

Text #4

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0487. Min Act: -0.0001

Data Index: 752509 (C4 (Web Text))

Max Activating Token Index: 825

Text #5

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0481. Min Act: -0.0001

Data Index: 446474 (C4 (Web Text))

Max Activating Token Index: 673

Text #6

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0412. Min Act: -0.0001

Data Index: 78310 (C4 (Web Text))

Max Activating Token Index: 676

Text #7

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0426. Min Act: -0.0001

Data Index: 531446 (C4 (Web Text))

Max Activating Token Index: 272

Text #8

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0483. Min Act: -0.0001

Data Index: 415551 (C4 (Web Text))

Max Activating Token Index: 768

Text #9

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0467. Min Act: -0.0001

Data Index: 1447736 (Python Code)

Max Activating Token Index: 816

Text #10

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0506. Min Act: -0.0001

Data Index: 934156 (C4 (Web Text))

Max Activating Token Index: 824

Text #11

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0439. Min Act: -0.0001

Data Index: 25572 (C4 (Web Text))

Max Activating Token Index: 603

Text #12

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0478. Min Act: -0.0001

Data Index: 323044 (C4 (Web Text))

Max Activating Token Index: 284

Text #13

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0453. Min Act: -0.0001

Data Index: 1612797 (Python Code)

Max Activating Token Index: 532

Text #14

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0451. Min Act: -0.0001

Data Index: 28776 (C4 (Web Text))

Max Activating Token Index: 471

Text #15

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0441. Min Act: -0.0001

Data Index: 573915 (C4 (Web Text))

Max Activating Token Index: 138

Text #16

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0360. Min Act: -0.0001

Data Index: 1077 (C4 (Web Text))

Max Activating Token Index: 602

Text #17

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0421. Min Act: -0.0001

Data Index: 1211927 (C4 (Web Text))

Max Activating Token Index: 262

Text #18

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0431. Min Act: -0.0001

Data Index: 1189578 (C4 (Web Text))

Max Activating Token Index: 457

Text #19

Max Range: 0.0515. Min Range: -0.0515

Max Act: 0.0452. Min Act: -0.0001

Data Index: 905496 (C4 (Web Text))

Max Activating Token Index: 392