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

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

Neuron 3912 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.5744. Min Act: -0.0001

Data Index: 1636003 (Python Code)

Max Activating Token Index: 906

Text #1

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.6211. Min Act: -0.0000

Data Index: 644525 (C4 (Web Text))

Max Activating Token Index: 607

Text #2

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.5816. Min Act: -0.0000

Data Index: 192746 (C4 (Web Text))

Max Activating Token Index: 510

Text #3

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.4238. Min Act: -0.0000

Data Index: 397066 (C4 (Web Text))

Max Activating Token Index: 916

Text #4

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.5424. Min Act: -0.0000

Data Index: 774434 (C4 (Web Text))

Max Activating Token Index: 944

Text #5

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.4124. Min Act: -0.0000

Data Index: 199256 (C4 (Web Text))

Max Activating Token Index: 567

Text #6

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.4681. Min Act: -0.0000

Data Index: 371299 (C4 (Web Text))

Max Activating Token Index: 271

Text #7

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.2636. Min Act: -0.0001

Data Index: 1493711 (Python Code)

Max Activating Token Index: 605

Text #8

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.2162. Min Act: -0.0000

Data Index: 299075 (C4 (Web Text))

Max Activating Token Index: 128

Text #9

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.2083. Min Act: -0.0000

Data Index: 359368 (C4 (Web Text))

Max Activating Token Index: 25

Text #10

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.3222. Min Act: -0.0001

Data Index: 1498924 (Python Code)

Max Activating Token Index: 892

Text #11

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.2885. Min Act: -0.0000

Data Index: 1648617 (Python Code)

Max Activating Token Index: 753

Text #12

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.1679. Min Act: -0.0000

Data Index: 973246 (C4 (Web Text))

Max Activating Token Index: 385

Text #13

Max Range: 2.6211. Min Range: -2.6211

Max Act: 1.8437. Min Act: -0.0001

Data Index: 1462517 (Python Code)

Max Activating Token Index: 743

Text #14

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.1583. Min Act: -0.0000

Data Index: 1286351 (C4 (Web Text))

Max Activating Token Index: 543

Text #15

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.1494. Min Act: -0.0000

Data Index: 535667 (C4 (Web Text))

Max Activating Token Index: 30

Text #16

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.1251. Min Act: -0.0000

Data Index: 1318938 (C4 (Web Text))

Max Activating Token Index: 849

Text #17

Max Range: 2.6211. Min Range: -2.6211

Max Act: 1.9046. Min Act: -0.0000

Data Index: 336389 (C4 (Web Text))

Max Activating Token Index: 575

Text #18

Max Range: 2.6211. Min Range: -2.6211

Max Act: 2.0740. Min Act: -0.0000

Data Index: 42653 (C4 (Web Text))

Max Activating Token Index: 32

Text #19

Max Range: 2.6211. Min Range: -2.6211

Max Act: 1.9528. Min Act: -0.0000

Data Index: 327829 (C4 (Web Text))

Max Activating Token Index: 563