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

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

Neuron 2560 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.1210. Min Range: -1.1210

Max Act: 1.1210. Min Act: -0.0000

Data Index: 998683 (C4 (Web Text))

Max Activating Token Index: 846

Text #1

Max Range: 1.1210. Min Range: -1.1210

Max Act: 1.0024. Min Act: -0.0000

Data Index: 396744 (C4 (Web Text))

Max Activating Token Index: 692

Text #2

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.9136. Min Act: -0.0000

Data Index: 94888 (C4 (Web Text))

Max Activating Token Index: 774

Text #3

Max Range: 1.1210. Min Range: -1.1210

Max Act: 1.0389. Min Act: -0.0000

Data Index: 1102112 (C4 (Web Text))

Max Activating Token Index: 642

Text #4

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.9052. Min Act: -0.0000

Data Index: 1576905 (Python Code)

Max Activating Token Index: 125

Text #5

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8759. Min Act: -0.0000

Data Index: 1568220 (Python Code)

Max Activating Token Index: 930

Text #6

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8727. Min Act: -0.0000

Data Index: 1377317 (Python Code)

Max Activating Token Index: 927

Text #7

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8281. Min Act: -0.0000

Data Index: 1365220 (Python Code)

Max Activating Token Index: 878

Text #8

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8290. Min Act: -0.0000

Data Index: 1491157 (Python Code)

Max Activating Token Index: 176

Text #9

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8408. Min Act: -0.0000

Data Index: 1640180 (Python Code)

Max Activating Token Index: 560

Text #10

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8181. Min Act: -0.0000

Data Index: 1354761 (C4 (Web Text))

Max Activating Token Index: 424

Text #11

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8062. Min Act: -0.0000

Data Index: 1656426 (Python Code)

Max Activating Token Index: 826

Text #12

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7722. Min Act: -0.0000

Data Index: 1538154 (Python Code)

Max Activating Token Index: 70

Text #13

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7641. Min Act: -0.0000

Data Index: 530983 (C4 (Web Text))

Max Activating Token Index: 1004

Text #14

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7384. Min Act: -0.0000

Data Index: 518514 (C4 (Web Text))

Max Activating Token Index: 785

Text #15

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7677. Min Act: -0.0000

Data Index: 1009753 (C4 (Web Text))

Max Activating Token Index: 817

Text #16

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7563. Min Act: -0.0000

Data Index: 1477895 (Python Code)

Max Activating Token Index: 1003

Text #17

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.8325. Min Act: -0.0000

Data Index: 1438941 (Python Code)

Max Activating Token Index: 861

Text #18

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7467. Min Act: -0.0000

Data Index: 1592847 (Python Code)

Max Activating Token Index: 187

Text #19

Max Range: 1.1210. Min Range: -1.1210

Max Act: 0.7826. Min Act: -0.0000

Data Index: 1472999 (Python Code)

Max Activating Token Index: 749