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

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

Neuron 1991 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: 0.0574. Min Range: -0.0574

Max Act: 0.0574. Min Act: -0.0005

Data Index: 518199 (C4 (Web Text))

Max Activating Token Index: 352

Text #1

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0573. Min Act: -0.0005

Data Index: 240484 (C4 (Web Text))

Max Activating Token Index: 371

Text #2

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0517. Min Act: -0.0005

Data Index: 1296424 (C4 (Web Text))

Max Activating Token Index: 834

Text #3

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0534. Min Act: -0.0005

Data Index: 1380897 (Python Code)

Max Activating Token Index: 759

Text #4

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0514. Min Act: -0.0005

Data Index: 1606888 (Python Code)

Max Activating Token Index: 114

Text #5

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0508. Min Act: -0.0005

Data Index: 1625665 (Python Code)

Max Activating Token Index: 746

Text #6

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0481. Min Act: -0.0005

Data Index: 659383 (C4 (Web Text))

Max Activating Token Index: 634

Text #7

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0517. Min Act: -0.0005

Data Index: 688998 (C4 (Web Text))

Max Activating Token Index: 38

Text #8

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0504. Min Act: -0.0005

Data Index: 564413 (C4 (Web Text))

Max Activating Token Index: 553

Text #9

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0520. Min Act: -0.0004

Data Index: 1255110 (C4 (Web Text))

Max Activating Token Index: 961

Text #10

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0499. Min Act: -0.0005

Data Index: 455657 (C4 (Web Text))

Max Activating Token Index: 194

Text #11

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0511. Min Act: -0.0005

Data Index: 272719 (C4 (Web Text))

Max Activating Token Index: 961

Text #12

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0507. Min Act: -0.0005

Data Index: 1489314 (Python Code)

Max Activating Token Index: 840

Text #13

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0497. Min Act: -0.0005

Data Index: 81344 (C4 (Web Text))

Max Activating Token Index: 727

Text #14

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0503. Min Act: -0.0005

Data Index: 1130466 (C4 (Web Text))

Max Activating Token Index: 961

Text #15

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0491. Min Act: -0.0005

Data Index: 1442810 (Python Code)

Max Activating Token Index: 451

Text #16

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0469. Min Act: -0.0005

Data Index: 398228 (C4 (Web Text))

Max Activating Token Index: 83

Text #17

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0526. Min Act: -0.0005

Data Index: 1456638 (Python Code)

Max Activating Token Index: 884

Text #18

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0477. Min Act: -0.0005

Data Index: 1221475 (C4 (Web Text))

Max Activating Token Index: 88

Text #19

Max Range: 0.0574. Min Range: -0.0574

Max Act: 0.0493. Min Act: -0.0005

Data Index: 99430 (C4 (Web Text))

Max Activating Token Index: 593