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

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

Neuron 3199 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: 1.7095. Min Range: -1.7095

Max Act: 1.5543. Min Act: -0.0000

Data Index: 1583060 (Python Code)

Max Activating Token Index: 866

Text #1

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.7095. Min Act: -0.0000

Data Index: 1023060 (C4 (Web Text))

Max Activating Token Index: 898

Text #2

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.4375. Min Act: -0.0000

Data Index: 1203487 (C4 (Web Text))

Max Activating Token Index: 932

Text #3

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.4051. Min Act: -0.0000

Data Index: 1150237 (C4 (Web Text))

Max Activating Token Index: 794

Text #4

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.5266. Min Act: -0.0000

Data Index: 1614118 (Python Code)

Max Activating Token Index: 847

Text #5

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.3632. Min Act: -0.0000

Data Index: 149696 (C4 (Web Text))

Max Activating Token Index: 266

Text #6

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.4478. Min Act: -0.0000

Data Index: 1648782 (Python Code)

Max Activating Token Index: 904

Text #7

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.5058. Min Act: -0.0000

Data Index: 171311 (C4 (Web Text))

Max Activating Token Index: 903

Text #8

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.3854. Min Act: -0.0000

Data Index: 1446717 (Python Code)

Max Activating Token Index: 621

Text #9

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.3782. Min Act: -0.0000

Data Index: 1607009 (Python Code)

Max Activating Token Index: 53

Text #10

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.3346. Min Act: -0.0000

Data Index: 1254012 (C4 (Web Text))

Max Activating Token Index: 1009

Text #11

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.3533. Min Act: -0.0000

Data Index: 1180206 (C4 (Web Text))

Max Activating Token Index: 360

Text #12

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2043. Min Act: -0.0000

Data Index: 1229966 (C4 (Web Text))

Max Activating Token Index: 887

Text #13

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2377. Min Act: -0.0000

Data Index: 1211415 (C4 (Web Text))

Max Activating Token Index: 661

Text #14

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2327. Min Act: -0.0000

Data Index: 1559234 (Python Code)

Max Activating Token Index: 566

Text #15

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.1547. Min Act: -0.0000

Data Index: 145702 (C4 (Web Text))

Max Activating Token Index: 349

Text #16

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2120. Min Act: -0.0000

Data Index: 793890 (C4 (Web Text))

Max Activating Token Index: 661

Text #17

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2125. Min Act: -0.0000

Data Index: 1117539 (C4 (Web Text))

Max Activating Token Index: 488

Text #18

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.2539. Min Act: -0.0000

Data Index: 227673 (C4 (Web Text))

Max Activating Token Index: 855

Text #19

Max Range: 1.7095. Min Range: -1.7095

Max Act: 1.1934. Min Act: -0.0000

Data Index: 1532965 (Python Code)

Max Activating Token Index: 735