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

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

Neuron 2867 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.0326. Min Range: -1.0326

Max Act: 1.0201. Min Act: -0.0000

Data Index: 251 (C4 (Web Text))

Max Activating Token Index: 697

Text #1

Max Range: 1.0326. Min Range: -1.0326

Max Act: 1.0326. Min Act: -0.0000

Data Index: 486343 (C4 (Web Text))

Max Activating Token Index: 595

Text #2

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9671. Min Act: -0.0000

Data Index: 1269420 (C4 (Web Text))

Max Activating Token Index: 260

Text #3

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9733. Min Act: -0.0000

Data Index: 1569981 (Python Code)

Max Activating Token Index: 17

Text #4

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9911. Min Act: -0.0000

Data Index: 1519519 (Python Code)

Max Activating Token Index: 904

Text #5

Max Range: 1.0326. Min Range: -1.0326

Max Act: 1.0114. Min Act: -0.0000

Data Index: 292927 (C4 (Web Text))

Max Activating Token Index: 398

Text #6

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9909. Min Act: -0.0000

Data Index: 1355620 (C4 (Web Text))

Max Activating Token Index: 757

Text #7

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9595. Min Act: -0.0000

Data Index: 122585 (C4 (Web Text))

Max Activating Token Index: 605

Text #8

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9211. Min Act: -0.0000

Data Index: 955928 (C4 (Web Text))

Max Activating Token Index: 338

Text #9

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9076. Min Act: -0.0000

Data Index: 101818 (C4 (Web Text))

Max Activating Token Index: 690

Text #10

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9201. Min Act: -0.0000

Data Index: 57233 (C4 (Web Text))

Max Activating Token Index: 667

Text #11

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9225. Min Act: -0.0000

Data Index: 920142 (C4 (Web Text))

Max Activating Token Index: 999

Text #12

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9097. Min Act: -0.0000

Data Index: 1362515 (Python Code)

Max Activating Token Index: 982

Text #13

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.9234. Min Act: -0.0000

Data Index: 867498 (C4 (Web Text))

Max Activating Token Index: 430

Text #14

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8331. Min Act: -0.0000

Data Index: 1649115 (Python Code)

Max Activating Token Index: 94

Text #15

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8796. Min Act: -0.0000

Data Index: 1252415 (C4 (Web Text))

Max Activating Token Index: 42

Text #16

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8537. Min Act: -0.0000

Data Index: 68493 (C4 (Web Text))

Max Activating Token Index: 204

Text #17

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8428. Min Act: -0.0000

Data Index: 1620396 (Python Code)

Max Activating Token Index: 822

Text #18

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8578. Min Act: -0.0000

Data Index: 1091735 (C4 (Web Text))

Max Activating Token Index: 277

Text #19

Max Range: 1.0326. Min Range: -1.0326

Max Act: 0.8855. Min Act: -0.0000

Data Index: 741993 (C4 (Web Text))

Max Activating Token Index: 104