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

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

Neuron 3646 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: 1.0959. Min Range: -1.0959

Max Act: 1.0166. Min Act: -0.0001

Data Index: 874445 (C4 (Web Text))

Max Activating Token Index: 86

Text #1

Max Range: 1.0959. Min Range: -1.0959

Max Act: 1.0959. Min Act: -0.0001

Data Index: 1074174 (C4 (Web Text))

Max Activating Token Index: 231

Text #2

Max Range: 1.0959. Min Range: -1.0959

Max Act: 1.0098. Min Act: -0.0001

Data Index: 207344 (C4 (Web Text))

Max Activating Token Index: 28

Text #3

Max Range: 1.0959. Min Range: -1.0959

Max Act: 1.0148. Min Act: -0.0001

Data Index: 61877 (C4 (Web Text))

Max Activating Token Index: 832

Text #4

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9679. Min Act: -0.0001

Data Index: 1574426 (Python Code)

Max Activating Token Index: 455

Text #5

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9836. Min Act: -0.0001

Data Index: 469910 (C4 (Web Text))

Max Activating Token Index: 88

Text #6

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9370. Min Act: -0.0001

Data Index: 1388593 (Python Code)

Max Activating Token Index: 31

Text #7

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9140. Min Act: -0.0001

Data Index: 1620178 (Python Code)

Max Activating Token Index: 664

Text #8

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8977. Min Act: -0.0001

Data Index: 210837 (C4 (Web Text))

Max Activating Token Index: 131

Text #9

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9359. Min Act: -0.0001

Data Index: 1403418 (Python Code)

Max Activating Token Index: 775

Text #10

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9609. Min Act: -0.0001

Data Index: 1295002 (C4 (Web Text))

Max Activating Token Index: 337

Text #11

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8859. Min Act: -0.0001

Data Index: 1446955 (Python Code)

Max Activating Token Index: 95

Text #12

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8843. Min Act: -0.0001

Data Index: 796879 (C4 (Web Text))

Max Activating Token Index: 46

Text #13

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8842. Min Act: -0.0001

Data Index: 1570734 (Python Code)

Max Activating Token Index: 552

Text #14

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9104. Min Act: -0.0001

Data Index: 883589 (C4 (Web Text))

Max Activating Token Index: 19

Text #15

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8803. Min Act: -0.0001

Data Index: 1445543 (Python Code)

Max Activating Token Index: 320

Text #16

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9057. Min Act: -0.0001

Data Index: 931294 (C4 (Web Text))

Max Activating Token Index: 136

Text #17

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.9085. Min Act: -0.0001

Data Index: 1556413 (Python Code)

Max Activating Token Index: 7

Text #18

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8952. Min Act: -0.0001

Data Index: 23344 (C4 (Web Text))

Max Activating Token Index: 29

Text #19

Max Range: 1.0959. Min Range: -1.0959

Max Act: 0.8890. Min Act: -0.0001

Data Index: 851171 (C4 (Web Text))

Max Activating Token Index: 5