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

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

Neuron 575 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.9447. Min Act: -0.1700

Data Index: 288423 (C4 (Web Text))

Max Activating Token Index: 865

Text #1

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.2260. Min Act: -0.1700

Data Index: 133278 (C4 (Web Text))

Max Activating Token Index: 664

Text #2

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.1322. Min Act: -0.1700

Data Index: 832389 (C4 (Web Text))

Max Activating Token Index: 504

Text #3

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.1010. Min Act: -0.1700

Data Index: 829867 (C4 (Web Text))

Max Activating Token Index: 822

Text #4

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.1010. Min Act: -0.1700

Data Index: 1506968 (Python Code)

Max Activating Token Index: 963

Text #5

Max Range: 7.9447. Min Range: -7.9447

Max Act: 7.0072. Min Act: -0.1700

Data Index: 870586 (C4 (Web Text))

Max Activating Token Index: 550

Text #6

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.8822. Min Act: -0.1700

Data Index: 629063 (C4 (Web Text))

Max Activating Token Index: 971

Text #7

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.9135. Min Act: -0.1700

Data Index: 1518161 (Python Code)

Max Activating Token Index: 982

Text #8

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.9447. Min Act: -0.1700

Data Index: 1484239 (Python Code)

Max Activating Token Index: 837

Text #9

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.9135. Min Act: -0.1700

Data Index: 816303 (C4 (Web Text))

Max Activating Token Index: 556

Text #10

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7885. Min Act: -0.1700

Data Index: 400024 (C4 (Web Text))

Max Activating Token Index: 339

Text #11

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7572. Min Act: -0.1700

Data Index: 113899 (C4 (Web Text))

Max Activating Token Index: 523

Text #12

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7885. Min Act: -0.1700

Data Index: 100174 (C4 (Web Text))

Max Activating Token Index: 993

Text #13

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7572. Min Act: -0.1700

Data Index: 786131 (C4 (Web Text))

Max Activating Token Index: 593

Text #14

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.6010. Min Act: -0.1700

Data Index: 1566065 (Python Code)

Max Activating Token Index: 826

Text #15

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7260. Min Act: -0.1700

Data Index: 1428106 (Python Code)

Max Activating Token Index: 615

Text #16

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.7885. Min Act: -0.1700

Data Index: 1362337 (Python Code)

Max Activating Token Index: 977

Text #17

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.6322. Min Act: -0.1700

Data Index: 1505014 (Python Code)

Max Activating Token Index: 610

Text #18

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.6635. Min Act: -0.1700

Data Index: 1328796 (C4 (Web Text))

Max Activating Token Index: 573

Text #19

Max Range: 7.9447. Min Range: -7.9447

Max Act: 6.6947. Min Act: -0.1700

Data Index: 1386170 (Python Code)

Max Activating Token Index: 845