Model: SoLU Model: 2 Layers, 2048 Neurons per Layer

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

Neuron 4 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-2l')

Text #0

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.3056. Min Act: -0.0003

Data Index: 1268835 (C4 (Web Text))

Max Activating Token Index: 461

Text #1

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2834. Min Act: -0.0003

Data Index: 280356 (C4 (Web Text))

Max Activating Token Index: 434

Text #2

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2842. Min Act: -0.0003

Data Index: 924718 (C4 (Web Text))

Max Activating Token Index: 475

Text #3

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2448. Min Act: -0.0003

Data Index: 1042871 (C4 (Web Text))

Max Activating Token Index: 993

Text #4

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2382. Min Act: -0.0003

Data Index: 829933 (C4 (Web Text))

Max Activating Token Index: 740

Text #5

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2308. Min Act: -0.0003

Data Index: 1006264 (C4 (Web Text))

Max Activating Token Index: 808

Text #6

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2264. Min Act: -0.0003

Data Index: 107418 (C4 (Web Text))

Max Activating Token Index: 154

Text #7

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2237. Min Act: -0.0003

Data Index: 626489 (C4 (Web Text))

Max Activating Token Index: 976

Text #8

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2211. Min Act: -0.0003

Data Index: 45387 (C4 (Web Text))

Max Activating Token Index: 846

Text #9

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2187. Min Act: -0.0003

Data Index: 1532880 (Python Code)

Max Activating Token Index: 1019

Text #10

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2177. Min Act: -0.0004

Data Index: 952628 (C4 (Web Text))

Max Activating Token Index: 992

Text #11

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2102. Min Act: -0.0004

Data Index: 1032246 (C4 (Web Text))

Max Activating Token Index: 921

Text #12

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2061. Min Act: -0.0004

Data Index: 821817 (C4 (Web Text))

Max Activating Token Index: 951

Text #13

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2038. Min Act: -0.0003

Data Index: 979463 (C4 (Web Text))

Max Activating Token Index: 827

Text #14

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2170. Min Act: -0.0003

Data Index: 1429619 (Python Code)

Max Activating Token Index: 983

Text #15

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2105. Min Act: -0.0003

Data Index: 121114 (C4 (Web Text))

Max Activating Token Index: 670

Text #16

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2087. Min Act: -0.0003

Data Index: 395680 (C4 (Web Text))

Max Activating Token Index: 1007

Text #17

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2100. Min Act: -0.0003

Data Index: 1198755 (C4 (Web Text))

Max Activating Token Index: 909

Text #18

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2159. Min Act: -0.0003

Data Index: 752071 (C4 (Web Text))

Max Activating Token Index: 919

Text #19

Max Range: 0.3056. Min Range: -0.3056

Max Act: 0.2071. Min Act: -0.0003

Data Index: 1053868 (C4 (Web Text))

Max Activating Token Index: 907