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

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

Neuron 4586 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 13.6962. Min Range: -13.6962

Max Act: 13.6962. Min Act: -0.0001

Data Index: 1454678 (Python Code)

Max Activating Token Index: 941

Text #1

Max Range: 13.6962. Min Range: -13.6962

Max Act: 11.8460. Min Act: -0.0001

Data Index: 588021 (C4 (Web Text))

Max Activating Token Index: 747

Text #2

Max Range: 13.6962. Min Range: -13.6962

Max Act: 11.1511. Min Act: -0.0001

Data Index: 326456 (C4 (Web Text))

Max Activating Token Index: 655

Text #3

Max Range: 13.6962. Min Range: -13.6962

Max Act: 10.9463. Min Act: -0.0001

Data Index: 1405119 (Python Code)

Max Activating Token Index: 953

Text #4

Max Range: 13.6962. Min Range: -13.6962

Max Act: 11.2096. Min Act: -0.0001

Data Index: 1575070 (Python Code)

Max Activating Token Index: 826

Text #5

Max Range: 13.6962. Min Range: -13.6962

Max Act: 9.2097. Min Act: -0.0001

Data Index: 1405737 (Python Code)

Max Activating Token Index: 1023

Text #6

Max Range: 13.6962. Min Range: -13.6962

Max Act: 7.8332. Min Act: -0.0001

Data Index: 1348845 (C4 (Web Text))

Max Activating Token Index: 827

Text #7

Max Range: 13.6962. Min Range: -13.6962

Max Act: 8.6514. Min Act: -0.0001

Data Index: 1371544 (Python Code)

Max Activating Token Index: 59

Text #8

Max Range: 13.6962. Min Range: -13.6962

Max Act: 4.0264. Min Act: -0.0001

Data Index: 1010886 (C4 (Web Text))

Max Activating Token Index: 616

Text #9

Max Range: 13.6962. Min Range: -13.6962

Max Act: 5.1176. Min Act: -0.0001

Data Index: 172178 (C4 (Web Text))

Max Activating Token Index: 626

Text #10

Max Range: 13.6962. Min Range: -13.6962

Max Act: 5.3111. Min Act: -0.0001

Data Index: 663584 (C4 (Web Text))

Max Activating Token Index: 86

Text #11

Max Range: 13.6962. Min Range: -13.6962

Max Act: 4.8293. Min Act: -0.0001

Data Index: 388108 (C4 (Web Text))

Max Activating Token Index: 768

Text #12

Max Range: 13.6962. Min Range: -13.6962

Max Act: 4.4295. Min Act: -0.0001

Data Index: 1004838 (C4 (Web Text))

Max Activating Token Index: 1023

Text #13

Max Range: 13.6962. Min Range: -13.6962

Max Act: 3.8705. Min Act: -0.0001

Data Index: 482569 (C4 (Web Text))

Max Activating Token Index: 573

Text #14

Max Range: 13.6962. Min Range: -13.6962

Max Act: 3.9465. Min Act: -0.0001

Data Index: 1540841 (Python Code)

Max Activating Token Index: 1023

Text #15

Max Range: 13.6962. Min Range: -13.6962

Max Act: 2.6601. Min Act: -0.0001

Data Index: 1008227 (C4 (Web Text))

Max Activating Token Index: 207

Text #16

Max Range: 13.6962. Min Range: -13.6962

Max Act: 3.8522. Min Act: -0.0001

Data Index: 1350634 (C4 (Web Text))

Max Activating Token Index: 30

Text #17

Max Range: 13.6962. Min Range: -13.6962

Max Act: 1.7303. Min Act: -0.0001

Data Index: 1526280 (Python Code)

Max Activating Token Index: 899

Text #18

Max Range: 13.6962. Min Range: -13.6962

Max Act: 2.3848. Min Act: -0.0001

Data Index: 101308 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 13.6962. Min Range: -13.6962

Max Act: 1.8379. Min Act: -0.0001

Data Index: 230119 (C4 (Web Text))

Max Activating Token Index: 694