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

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

Neuron 4185 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.7067. Min Act: -0.0000

Data Index: 927225 (C4 (Web Text))

Max Activating Token Index: 756

Text #1

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.2440. Min Act: -0.0000

Data Index: 192839 (C4 (Web Text))

Max Activating Token Index: 930

Text #2

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.1353. Min Act: -0.0000

Data Index: 88011 (C4 (Web Text))

Max Activating Token Index: 418

Text #3

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.1094. Min Act: -0.0000

Data Index: 1240347 (C4 (Web Text))

Max Activating Token Index: 724

Text #4

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.0645. Min Act: -0.0000

Data Index: 667613 (C4 (Web Text))

Max Activating Token Index: 23

Text #5

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.0575. Min Act: -0.0000

Data Index: 206249 (C4 (Web Text))

Max Activating Token Index: 55

Text #6

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.9878. Min Act: -0.0000

Data Index: 811144 (C4 (Web Text))

Max Activating Token Index: 102

Text #7

Max Range: 2.7067. Min Range: -2.7067

Max Act: 2.0812. Min Act: -0.0000

Data Index: 67576 (C4 (Web Text))

Max Activating Token Index: 162

Text #8

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.9630. Min Act: -0.0000

Data Index: 958906 (C4 (Web Text))

Max Activating Token Index: 920

Text #9

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.8215. Min Act: -0.0000

Data Index: 92375 (C4 (Web Text))

Max Activating Token Index: 245

Text #10

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.9162. Min Act: -0.0000

Data Index: 591096 (C4 (Web Text))

Max Activating Token Index: 983

Text #11

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.8847. Min Act: -0.0000

Data Index: 378707 (C4 (Web Text))

Max Activating Token Index: 928

Text #12

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.8932. Min Act: -0.0000

Data Index: 494247 (C4 (Web Text))

Max Activating Token Index: 62

Text #13

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.6994. Min Act: -0.0000

Data Index: 250494 (C4 (Web Text))

Max Activating Token Index: 976

Text #14

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.8514. Min Act: -0.0000

Data Index: 1359923 (Python Code)

Max Activating Token Index: 80

Text #15

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.7506. Min Act: -0.0000

Data Index: 628989 (C4 (Web Text))

Max Activating Token Index: 615

Text #16

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.7479. Min Act: -0.0000

Data Index: 1550483 (Python Code)

Max Activating Token Index: 604

Text #17

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.6200. Min Act: -0.0000

Data Index: 361540 (C4 (Web Text))

Max Activating Token Index: 88

Text #18

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.9190. Min Act: -0.0000

Data Index: 1526547 (Python Code)

Max Activating Token Index: 939

Text #19

Max Range: 2.7067. Min Range: -2.7067

Max Act: 1.8089. Min Act: -0.0000

Data Index: 159518 (C4 (Web Text))

Max Activating Token Index: 563