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

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

Neuron 4270 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1505. Min Act: -0.0001

Data Index: 903995 (C4 (Web Text))

Max Activating Token Index: 116

Text #1

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1536. Min Act: -0.0001

Data Index: 1594304 (Python Code)

Max Activating Token Index: 297

Text #2

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1482. Min Act: -0.0001

Data Index: 1632488 (Python Code)

Max Activating Token Index: 898

Text #3

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1415. Min Act: -0.0001

Data Index: 1571442 (Python Code)

Max Activating Token Index: 279

Text #4

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1341. Min Act: -0.0001

Data Index: 1400877 (Python Code)

Max Activating Token Index: 691

Text #5

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1308. Min Act: -0.0001

Data Index: 1345994 (C4 (Web Text))

Max Activating Token Index: 829

Text #6

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1246. Min Act: -0.0001

Data Index: 1584061 (Python Code)

Max Activating Token Index: 105

Text #7

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1117. Min Act: -0.0001

Data Index: 328423 (C4 (Web Text))

Max Activating Token Index: 757

Text #8

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1041. Min Act: -0.0001

Data Index: 84834 (C4 (Web Text))

Max Activating Token Index: 546

Text #9

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1058. Min Act: -0.0001

Data Index: 1485892 (Python Code)

Max Activating Token Index: 805

Text #10

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0966. Min Act: -0.0001

Data Index: 497882 (C4 (Web Text))

Max Activating Token Index: 532

Text #11

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1080. Min Act: -0.0001

Data Index: 1529599 (Python Code)

Max Activating Token Index: 293

Text #12

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0988. Min Act: -0.0001

Data Index: 2746 (C4 (Web Text))

Max Activating Token Index: 200

Text #13

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.1043. Min Act: -0.0001

Data Index: 5393 (C4 (Web Text))

Max Activating Token Index: 741

Text #14

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0994. Min Act: -0.0001

Data Index: 1322788 (C4 (Web Text))

Max Activating Token Index: 202

Text #15

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0966. Min Act: -0.0001

Data Index: 1491826 (Python Code)

Max Activating Token Index: 724

Text #16

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0941. Min Act: -0.0001

Data Index: 74134 (C4 (Web Text))

Max Activating Token Index: 733

Text #17

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0870. Min Act: -0.0001

Data Index: 1519743 (Python Code)

Max Activating Token Index: 932

Text #18

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0879. Min Act: -0.0001

Data Index: 1564927 (Python Code)

Max Activating Token Index: 932

Text #19

Max Range: 0.1536. Min Range: -0.1536

Max Act: 0.0890. Min Act: -0.0001

Data Index: 1644115 (Python Code)

Max Activating Token Index: 145