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

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

Neuron 4670 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: 2.8769. Min Range: -2.8769

Max Act: 2.8769. Min Act: -0.0000

Data Index: 195819 (C4 (Web Text))

Max Activating Token Index: 134

Text #1

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.2881. Min Act: -0.0000

Data Index: 671612 (C4 (Web Text))

Max Activating Token Index: 592

Text #2

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.2434. Min Act: -0.0000

Data Index: 168843 (C4 (Web Text))

Max Activating Token Index: 895

Text #3

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.1906. Min Act: -0.0000

Data Index: 490668 (C4 (Web Text))

Max Activating Token Index: 477

Text #4

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.0817. Min Act: -0.0000

Data Index: 469770 (C4 (Web Text))

Max Activating Token Index: 870

Text #5

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.2021. Min Act: -0.0000

Data Index: 1025045 (C4 (Web Text))

Max Activating Token Index: 407

Text #6

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.0735. Min Act: -0.0000

Data Index: 1342011 (C4 (Web Text))

Max Activating Token Index: 858

Text #7

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.0597. Min Act: -0.0000

Data Index: 784883 (C4 (Web Text))

Max Activating Token Index: 381

Text #8

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.0449. Min Act: -0.0000

Data Index: 1580509 (Python Code)

Max Activating Token Index: 907

Text #9

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8984. Min Act: -0.0000

Data Index: 588040 (C4 (Web Text))

Max Activating Token Index: 325

Text #10

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8906. Min Act: -0.0000

Data Index: 1556339 (Python Code)

Max Activating Token Index: 631

Text #11

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.9828. Min Act: -0.0000

Data Index: 1227979 (C4 (Web Text))

Max Activating Token Index: 720

Text #12

Max Range: 2.8769. Min Range: -2.8769

Max Act: 2.0337. Min Act: -0.0000

Data Index: 265500 (C4 (Web Text))

Max Activating Token Index: 621

Text #13

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8994. Min Act: -0.0000

Data Index: 653633 (C4 (Web Text))

Max Activating Token Index: 354

Text #14

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8997. Min Act: -0.0000

Data Index: 127829 (C4 (Web Text))

Max Activating Token Index: 7

Text #15

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.9918. Min Act: -0.0000

Data Index: 320402 (C4 (Web Text))

Max Activating Token Index: 525

Text #16

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8917. Min Act: -0.0000

Data Index: 1415962 (Python Code)

Max Activating Token Index: 370

Text #17

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8787. Min Act: -0.0000

Data Index: 116579 (C4 (Web Text))

Max Activating Token Index: 408

Text #18

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8627. Min Act: -0.0000

Data Index: 383169 (C4 (Web Text))

Max Activating Token Index: 464

Text #19

Max Range: 2.8769. Min Range: -2.8769

Max Act: 1.8415. Min Act: -0.0000

Data Index: 667318 (C4 (Web Text))

Max Activating Token Index: 96