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

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

Neuron 3732 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0800. Min Act: -0.0002

Data Index: 1649636 (Python Code)

Max Activating Token Index: 102

Text #1

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0810. Min Act: -0.0002

Data Index: 1124999 (C4 (Web Text))

Max Activating Token Index: 387

Text #2

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0740. Min Act: -0.0002

Data Index: 1441478 (Python Code)

Max Activating Token Index: 444

Text #3

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0730. Min Act: -0.0002

Data Index: 1063071 (C4 (Web Text))

Max Activating Token Index: 651

Text #4

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0741. Min Act: -0.0002

Data Index: 328625 (C4 (Web Text))

Max Activating Token Index: 909

Text #5

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0716. Min Act: -0.0002

Data Index: 617824 (C4 (Web Text))

Max Activating Token Index: 152

Text #6

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0741. Min Act: -0.0002

Data Index: 1346693 (C4 (Web Text))

Max Activating Token Index: 42

Text #7

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0709. Min Act: -0.0002

Data Index: 934684 (C4 (Web Text))

Max Activating Token Index: 54

Text #8

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0734. Min Act: -0.0002

Data Index: 1535674 (Python Code)

Max Activating Token Index: 7

Text #9

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0774. Min Act: -0.0002

Data Index: 1275953 (C4 (Web Text))

Max Activating Token Index: 165

Text #10

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0663. Min Act: -0.0002

Data Index: 391069 (C4 (Web Text))

Max Activating Token Index: 722

Text #11

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0737. Min Act: -0.0002

Data Index: 1558851 (Python Code)

Max Activating Token Index: 460

Text #12

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0753. Min Act: -0.0002

Data Index: 1524331 (Python Code)

Max Activating Token Index: 47

Text #13

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0680. Min Act: -0.0002

Data Index: 1181002 (C4 (Web Text))

Max Activating Token Index: 499

Text #14

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0751. Min Act: -0.0002

Data Index: 537680 (C4 (Web Text))

Max Activating Token Index: 226

Text #15

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0674. Min Act: -0.0002

Data Index: 49330 (C4 (Web Text))

Max Activating Token Index: 139

Text #16

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0655. Min Act: -0.0002

Data Index: 1510854 (Python Code)

Max Activating Token Index: 806

Text #17

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0692. Min Act: -0.0002

Data Index: 1419805 (Python Code)

Max Activating Token Index: 207

Text #18

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0665. Min Act: -0.0002

Data Index: 1119425 (C4 (Web Text))

Max Activating Token Index: 111

Text #19

Max Range: 0.0810. Min Range: -0.0810

Max Act: 0.0688. Min Act: -0.0002

Data Index: 60721 (C4 (Web Text))

Max Activating Token Index: 439