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

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

Neuron 3794 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: 0.7898. Min Range: -0.7898

Max Act: 0.7898. Min Act: -0.0001

Data Index: 1443304 (Python Code)

Max Activating Token Index: 300

Text #1

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6689. Min Act: -0.0000

Data Index: 1101708 (C4 (Web Text))

Max Activating Token Index: 959

Text #2

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.7674. Min Act: -0.0001

Data Index: 1419977 (Python Code)

Max Activating Token Index: 257

Text #3

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.7008. Min Act: -0.0000

Data Index: 431241 (C4 (Web Text))

Max Activating Token Index: 24

Text #4

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6775. Min Act: -0.0001

Data Index: 1435318 (Python Code)

Max Activating Token Index: 763

Text #5

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6797. Min Act: -0.0000

Data Index: 165611 (C4 (Web Text))

Max Activating Token Index: 391

Text #6

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6737. Min Act: -0.0000

Data Index: 1570790 (Python Code)

Max Activating Token Index: 407

Text #7

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6924. Min Act: -0.0001

Data Index: 1564707 (Python Code)

Max Activating Token Index: 141

Text #8

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5968. Min Act: -0.0000

Data Index: 1255305 (C4 (Web Text))

Max Activating Token Index: 451

Text #9

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.6507. Min Act: -0.0000

Data Index: 601546 (C4 (Web Text))

Max Activating Token Index: 116

Text #10

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5360. Min Act: -0.0001

Data Index: 1387330 (Python Code)

Max Activating Token Index: 21

Text #11

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5863. Min Act: -0.0001

Data Index: 1148254 (C4 (Web Text))

Max Activating Token Index: 966

Text #12

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5429. Min Act: -0.0001

Data Index: 1142368 (C4 (Web Text))

Max Activating Token Index: 547

Text #13

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5311. Min Act: -0.0001

Data Index: 717251 (C4 (Web Text))

Max Activating Token Index: 938

Text #14

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5604. Min Act: -0.0000

Data Index: 729500 (C4 (Web Text))

Max Activating Token Index: 324

Text #15

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5261. Min Act: -0.0000

Data Index: 1641658 (Python Code)

Max Activating Token Index: 286

Text #16

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5157. Min Act: -0.0000

Data Index: 1216838 (C4 (Web Text))

Max Activating Token Index: 97

Text #17

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5345. Min Act: -0.0000

Data Index: 1490444 (Python Code)

Max Activating Token Index: 36

Text #18

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5115. Min Act: -0.0000

Data Index: 38143 (C4 (Web Text))

Max Activating Token Index: 318

Text #19

Max Range: 0.7898. Min Range: -0.7898

Max Act: 0.5193. Min Act: -0.0001

Data Index: 1596280 (Python Code)

Max Activating Token Index: 572