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

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

Neuron 2605 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: 0.0987. Min Range: -0.0987

Max Act: 0.0987. Min Act: -0.0001

Data Index: 72491 (C4 (Web Text))

Max Activating Token Index: 560

Text #1

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0980. Min Act: -0.0001

Data Index: 1488532 (Python Code)

Max Activating Token Index: 938

Text #2

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0973. Min Act: -0.0001

Data Index: 917869 (C4 (Web Text))

Max Activating Token Index: 887

Text #3

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0966. Min Act: -0.0001

Data Index: 788092 (C4 (Web Text))

Max Activating Token Index: 598

Text #4

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0875. Min Act: -0.0001

Data Index: 1122802 (C4 (Web Text))

Max Activating Token Index: 80

Text #5

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0938. Min Act: -0.0001

Data Index: 227731 (C4 (Web Text))

Max Activating Token Index: 451

Text #6

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0971. Min Act: -0.0001

Data Index: 643542 (C4 (Web Text))

Max Activating Token Index: 574

Text #7

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0908. Min Act: -0.0001

Data Index: 1526072 (Python Code)

Max Activating Token Index: 951

Text #8

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0907. Min Act: -0.0001

Data Index: 141352 (C4 (Web Text))

Max Activating Token Index: 282

Text #9

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0842. Min Act: -0.0001

Data Index: 65836 (C4 (Web Text))

Max Activating Token Index: 993

Text #10

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0932. Min Act: -0.0001

Data Index: 257950 (C4 (Web Text))

Max Activating Token Index: 647

Text #11

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0888. Min Act: -0.0001

Data Index: 87223 (C4 (Web Text))

Max Activating Token Index: 423

Text #12

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0806. Min Act: -0.0001

Data Index: 465821 (C4 (Web Text))

Max Activating Token Index: 728

Text #13

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0850. Min Act: -0.0001

Data Index: 1581092 (Python Code)

Max Activating Token Index: 145

Text #14

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0879. Min Act: -0.0001

Data Index: 647126 (C4 (Web Text))

Max Activating Token Index: 662

Text #15

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0777. Min Act: -0.0001

Data Index: 1025424 (C4 (Web Text))

Max Activating Token Index: 803

Text #16

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0824. Min Act: -0.0001

Data Index: 1542078 (Python Code)

Max Activating Token Index: 64

Text #17

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0814. Min Act: -0.0001

Data Index: 22184 (C4 (Web Text))

Max Activating Token Index: 858

Text #18

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0821. Min Act: -0.0001

Data Index: 1589629 (Python Code)

Max Activating Token Index: 273

Text #19

Max Range: 0.0987. Min Range: -0.0987

Max Act: 0.0815. Min Act: -0.0001

Data Index: 341507 (C4 (Web Text))

Max Activating Token Index: 153