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

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

Neuron 672 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1216. Min Act: -0.0000

Data Index: 1217549 (C4 (Web Text))

Max Activating Token Index: 18

Text #1

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1164. Min Act: -0.0000

Data Index: 1318559 (C4 (Web Text))

Max Activating Token Index: 15

Text #2

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1223. Min Act: -0.0000

Data Index: 1047596 (C4 (Web Text))

Max Activating Token Index: 120

Text #3

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1169. Min Act: -0.0000

Data Index: 260555 (C4 (Web Text))

Max Activating Token Index: 30

Text #4

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1092. Min Act: -0.0000

Data Index: 249621 (C4 (Web Text))

Max Activating Token Index: 69

Text #5

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1104. Min Act: -0.0000

Data Index: 1578118 (Python Code)

Max Activating Token Index: 536

Text #6

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1093. Min Act: -0.0000

Data Index: 1030230 (C4 (Web Text))

Max Activating Token Index: 68

Text #7

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1021. Min Act: -0.0000

Data Index: 1028053 (C4 (Web Text))

Max Activating Token Index: 699

Text #8

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1088. Min Act: -0.0000

Data Index: 191359 (C4 (Web Text))

Max Activating Token Index: 46

Text #9

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1028. Min Act: -0.0000

Data Index: 418553 (C4 (Web Text))

Max Activating Token Index: 21

Text #10

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0981. Min Act: -0.0000

Data Index: 984817 (C4 (Web Text))

Max Activating Token Index: 1011

Text #11

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1011. Min Act: -0.0000

Data Index: 661009 (C4 (Web Text))

Max Activating Token Index: 481

Text #12

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1011. Min Act: -0.0000

Data Index: 166285 (C4 (Web Text))

Max Activating Token Index: 254

Text #13

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.1040. Min Act: -0.0001

Data Index: 1480458 (Python Code)

Max Activating Token Index: 315

Text #14

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0976. Min Act: -0.0000

Data Index: 154934 (C4 (Web Text))

Max Activating Token Index: 752

Text #15

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0966. Min Act: -0.0000

Data Index: 426399 (C4 (Web Text))

Max Activating Token Index: 20

Text #16

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0999. Min Act: -0.0000

Data Index: 1175661 (C4 (Web Text))

Max Activating Token Index: 89

Text #17

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0963. Min Act: -0.0000

Data Index: 1297013 (C4 (Web Text))

Max Activating Token Index: 43

Text #18

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0956. Min Act: -0.0001

Data Index: 9310 (C4 (Web Text))

Max Activating Token Index: 35

Text #19

Max Range: 0.1223. Min Range: -0.1223

Max Act: 0.0957. Min Act: -0.0000

Data Index: 443576 (C4 (Web Text))

Max Activating Token Index: 170