Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2011 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1932. Min Act: -0.0001

Data Index: 785403 (C4 (Web Text))

Max Activating Token Index: 571

Text #1

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1946. Min Act: -0.0001

Data Index: 688526 (C4 (Web Text))

Max Activating Token Index: 103

Text #2

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.2076. Min Act: -0.0001

Data Index: 222731 (C4 (Web Text))

Max Activating Token Index: 789

Text #3

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1768. Min Act: -0.0000

Data Index: 1054126 (C4 (Web Text))

Max Activating Token Index: 918

Text #4

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1803. Min Act: -0.0001

Data Index: 1639851 (Python Code)

Max Activating Token Index: 97

Text #5

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1859. Min Act: -0.0000

Data Index: 1298553 (C4 (Web Text))

Max Activating Token Index: 574

Text #6

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1784. Min Act: -0.0001

Data Index: 1612513 (Python Code)

Max Activating Token Index: 942

Text #7

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1925. Min Act: -0.0001

Data Index: 1362042 (Python Code)

Max Activating Token Index: 584

Text #8

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1851. Min Act: -0.0001

Data Index: 1393409 (Python Code)

Max Activating Token Index: 116

Text #9

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1785. Min Act: -0.0001

Data Index: 1279111 (C4 (Web Text))

Max Activating Token Index: 811

Text #10

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1761. Min Act: -0.0001

Data Index: 615827 (C4 (Web Text))

Max Activating Token Index: 936

Text #11

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1805. Min Act: -0.0000

Data Index: 487826 (C4 (Web Text))

Max Activating Token Index: 514

Text #12

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1937. Min Act: -0.0000

Data Index: 1116147 (C4 (Web Text))

Max Activating Token Index: 739

Text #13

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1803. Min Act: -0.0000

Data Index: 1335843 (C4 (Web Text))

Max Activating Token Index: 704

Text #14

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1845. Min Act: -0.0001

Data Index: 440758 (C4 (Web Text))

Max Activating Token Index: 982

Text #15

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1580. Min Act: -0.0001

Data Index: 1479439 (Python Code)

Max Activating Token Index: 978

Text #16

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1542. Min Act: -0.0000

Data Index: 47670 (C4 (Web Text))

Max Activating Token Index: 932

Text #17

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1620. Min Act: -0.0001

Data Index: 873589 (C4 (Web Text))

Max Activating Token Index: 308

Text #18

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1687. Min Act: -0.0001

Data Index: 1600675 (Python Code)

Max Activating Token Index: 1003

Text #19

Max Range: 0.2076. Min Range: -0.2076

Max Act: 0.1673. Min Act: -0.0001

Data Index: 1052905 (C4 (Web Text))

Max Activating Token Index: 67