Model: GELU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 2046 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 4.1717. Min Range: -4.1717

Max Act: 4.1717. Min Act: -0.1700

Data Index: 1447045 (Python Code)

Max Activating Token Index: 829

Text #1

Max Range: 4.1717. Min Range: -4.1717

Max Act: 4.1404. Min Act: -0.1700

Data Index: 871569 (C4 (Web Text))

Max Activating Token Index: 677

Text #2

Max Range: 4.1717. Min Range: -4.1717

Max Act: 4.0467. Min Act: -0.1700

Data Index: 494435 (C4 (Web Text))

Max Activating Token Index: 948

Text #3

Max Range: 4.1717. Min Range: -4.1717

Max Act: 4.0467. Min Act: -0.1700

Data Index: 1491524 (Python Code)

Max Activating Token Index: 598

Text #4

Max Range: 4.1717. Min Range: -4.1717

Max Act: 4.0467. Min Act: -0.1700

Data Index: 1560374 (Python Code)

Max Activating Token Index: 81

Text #5

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9841. Min Act: -0.1700

Data Index: 1547647 (Python Code)

Max Activating Token Index: 35

Text #6

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9529. Min Act: -0.1700

Data Index: 642391 (C4 (Web Text))

Max Activating Token Index: 859

Text #7

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9529. Min Act: -0.1700

Data Index: 355003 (C4 (Web Text))

Max Activating Token Index: 815

Text #8

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9529. Min Act: -0.1700

Data Index: 1523185 (Python Code)

Max Activating Token Index: 479

Text #9

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9841. Min Act: -0.1700

Data Index: 1556621 (Python Code)

Max Activating Token Index: 498

Text #10

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9216. Min Act: -0.1700

Data Index: 1451952 (Python Code)

Max Activating Token Index: 135

Text #11

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9216. Min Act: -0.1700

Data Index: 1205272 (C4 (Web Text))

Max Activating Token Index: 941

Text #12

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8591. Min Act: -0.1700

Data Index: 1645640 (Python Code)

Max Activating Token Index: 760

Text #13

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.9529. Min Act: -0.1700

Data Index: 1425320 (Python Code)

Max Activating Token Index: 727

Text #14

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8903. Min Act: -0.1700

Data Index: 1606645 (Python Code)

Max Activating Token Index: 80

Text #15

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8903. Min Act: -0.1700

Data Index: 1371353 (Python Code)

Max Activating Token Index: 973

Text #16

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8903. Min Act: -0.1700

Data Index: 1613397 (Python Code)

Max Activating Token Index: 33

Text #17

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8591. Min Act: -0.1700

Data Index: 1557145 (Python Code)

Max Activating Token Index: 824

Text #18

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8591. Min Act: -0.1700

Data Index: 1555629 (Python Code)

Max Activating Token Index: 252

Text #19

Max Range: 4.1717. Min Range: -4.1717

Max Act: 3.8903. Min Act: -0.1700

Data Index: 1384674 (Python Code)

Max Activating Token Index: 6