Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 4000 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.4679. Min Act: -0.0001

Data Index: 561694 (C4 (Web Text))

Max Activating Token Index: 669

Text #1

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.9605. Min Act: -0.0001

Data Index: 494942 (C4 (Web Text))

Max Activating Token Index: 388

Text #2

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.2039. Min Act: -0.0000

Data Index: 428464 (C4 (Web Text))

Max Activating Token Index: 71

Text #3

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.2262. Min Act: -0.0001

Data Index: 866994 (C4 (Web Text))

Max Activating Token Index: 989

Text #4

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.2155. Min Act: -0.0001

Data Index: 431491 (C4 (Web Text))

Max Activating Token Index: 573

Text #5

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.3271. Min Act: -0.0001

Data Index: 1129209 (C4 (Web Text))

Max Activating Token Index: 669

Text #6

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.1965. Min Act: -0.0001

Data Index: 510365 (C4 (Web Text))

Max Activating Token Index: 369

Text #7

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.0254. Min Act: -0.0000

Data Index: 1068513 (C4 (Web Text))

Max Activating Token Index: 73

Text #8

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.0274. Min Act: -0.0001

Data Index: 978695 (C4 (Web Text))

Max Activating Token Index: 612

Text #9

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8172. Min Act: -0.0001

Data Index: 1000504 (C4 (Web Text))

Max Activating Token Index: 874

Text #10

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.2478. Min Act: -0.0001

Data Index: 864568 (C4 (Web Text))

Max Activating Token Index: 616

Text #11

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.0900. Min Act: -0.0001

Data Index: 1255627 (C4 (Web Text))

Max Activating Token Index: 456

Text #12

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8799. Min Act: -0.0001

Data Index: 545891 (C4 (Web Text))

Max Activating Token Index: 922

Text #13

Max Range: 3.4679. Min Range: -3.4679

Max Act: 3.1381. Min Act: -0.0001

Data Index: 180339 (C4 (Web Text))

Max Activating Token Index: 192

Text #14

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.9147. Min Act: -0.0001

Data Index: 1284652 (C4 (Web Text))

Max Activating Token Index: 868

Text #15

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.7694. Min Act: -0.0001

Data Index: 479823 (C4 (Web Text))

Max Activating Token Index: 89

Text #16

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8737. Min Act: -0.0000

Data Index: 1214323 (C4 (Web Text))

Max Activating Token Index: 568

Text #17

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8560. Min Act: -0.0001

Data Index: 258160 (C4 (Web Text))

Max Activating Token Index: 470

Text #18

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8484. Min Act: -0.0001

Data Index: 1323996 (C4 (Web Text))

Max Activating Token Index: 579

Text #19

Max Range: 3.4679. Min Range: -3.4679

Max Act: 2.8654. Min Act: -0.0001

Data Index: 461438 (C4 (Web Text))

Max Activating Token Index: 847