Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 614 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0611. Min Act: -0.0001

Data Index: 1391560 (Python Code)

Max Activating Token Index: 805

Text #1

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0683. Min Act: -0.0001

Data Index: 279232 (C4 (Web Text))

Max Activating Token Index: 943

Text #2

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0629. Min Act: -0.0001

Data Index: 82806 (C4 (Web Text))

Max Activating Token Index: 987

Text #3

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0606. Min Act: -0.0001

Data Index: 113034 (C4 (Web Text))

Max Activating Token Index: 16

Text #4

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0610. Min Act: -0.0001

Data Index: 654583 (C4 (Web Text))

Max Activating Token Index: 705

Text #5

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0634. Min Act: -0.0001

Data Index: 72489 (C4 (Web Text))

Max Activating Token Index: 655

Text #6

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0587. Min Act: -0.0001

Data Index: 1280116 (C4 (Web Text))

Max Activating Token Index: 719

Text #7

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0516. Min Act: -0.0001

Data Index: 1627355 (Python Code)

Max Activating Token Index: 849

Text #8

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0535. Min Act: -0.0001

Data Index: 216244 (C4 (Web Text))

Max Activating Token Index: 428

Text #9

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0526. Min Act: -0.0001

Data Index: 255582 (C4 (Web Text))

Max Activating Token Index: 14

Text #10

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0531. Min Act: -0.0001

Data Index: 760812 (C4 (Web Text))

Max Activating Token Index: 2

Text #11

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0531. Min Act: -0.0001

Data Index: 1236727 (C4 (Web Text))

Max Activating Token Index: 2

Text #12

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0508. Min Act: -0.0001

Data Index: 669210 (C4 (Web Text))

Max Activating Token Index: 90

Text #13

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0509. Min Act: -0.0001

Data Index: 1607864 (Python Code)

Max Activating Token Index: 783

Text #14

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0418. Min Act: -0.0001

Data Index: 711547 (C4 (Web Text))

Max Activating Token Index: 605

Text #15

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0503. Min Act: -0.0001

Data Index: 760869 (C4 (Web Text))

Max Activating Token Index: 337

Text #16

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0500. Min Act: -0.0001

Data Index: 504890 (C4 (Web Text))

Max Activating Token Index: 186

Text #17

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0514. Min Act: -0.0001

Data Index: 643497 (C4 (Web Text))

Max Activating Token Index: 820

Text #18

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0500. Min Act: -0.0001

Data Index: 469836 (C4 (Web Text))

Max Activating Token Index: 462

Text #19

Max Range: 0.0683. Min Range: -0.0683

Max Act: 0.0458. Min Act: -0.0001

Data Index: 618833 (C4 (Web Text))

Max Activating Token Index: 194