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

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

Neuron 361 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.8362. Min Act: 0.0000

Data Index: 1542033 (Python Code)

Max Activating Token Index: 389

Text #1

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7889. Min Act: -0.0000

Data Index: 837904 (C4 (Web Text))

Max Activating Token Index: 756

Text #2

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.8372. Min Act: 0.0000

Data Index: 1165179 (C4 (Web Text))

Max Activating Token Index: 355

Text #3

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7780. Min Act: -0.0000

Data Index: 511212 (C4 (Web Text))

Max Activating Token Index: 156

Text #4

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7994. Min Act: -0.0000

Data Index: 1158658 (C4 (Web Text))

Max Activating Token Index: 597

Text #5

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7664. Min Act: -0.0000

Data Index: 986495 (C4 (Web Text))

Max Activating Token Index: 978

Text #6

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7385. Min Act: 0.0000

Data Index: 1446982 (Python Code)

Max Activating Token Index: 6

Text #7

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.7058. Min Act: -0.0000

Data Index: 1342612 (C4 (Web Text))

Max Activating Token Index: 596

Text #8

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6937. Min Act: -0.0000

Data Index: 46333 (C4 (Web Text))

Max Activating Token Index: 382

Text #9

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6478. Min Act: -0.0000

Data Index: 1419379 (Python Code)

Max Activating Token Index: 939

Text #10

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6401. Min Act: -0.0000

Data Index: 899520 (C4 (Web Text))

Max Activating Token Index: 269

Text #11

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6721. Min Act: -0.0000

Data Index: 1468567 (Python Code)

Max Activating Token Index: 399

Text #12

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6176. Min Act: -0.0000

Data Index: 696544 (C4 (Web Text))

Max Activating Token Index: 900

Text #13

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6312. Min Act: -0.0000

Data Index: 1641072 (Python Code)

Max Activating Token Index: 181

Text #14

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6051. Min Act: 0.0001

Data Index: 1372668 (Python Code)

Max Activating Token Index: 346

Text #15

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6051. Min Act: -0.0000

Data Index: 1640976 (Python Code)

Max Activating Token Index: 92

Text #16

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6300. Min Act: 0.0000

Data Index: 86381 (C4 (Web Text))

Max Activating Token Index: 779

Text #17

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6100. Min Act: 0.0000

Data Index: 1141531 (C4 (Web Text))

Max Activating Token Index: 102

Text #18

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6011. Min Act: -0.0000

Data Index: 718031 (C4 (Web Text))

Max Activating Token Index: 256

Text #19

Max Range: 0.8372. Min Range: -0.8372

Max Act: 0.6352. Min Act: -0.0000

Data Index: 1210999 (C4 (Web Text))

Max Activating Token Index: 517