Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 575 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.7995. Min Act: -0.1700

Data Index: 670915 (C4 (Web Text))

Max Activating Token Index: 619

Text #1

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.7683. Min Act: -0.1700

Data Index: 132321 (C4 (Web Text))

Max Activating Token Index: 658

Text #2

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.7057. Min Act: -0.1700

Data Index: 177219 (C4 (Web Text))

Max Activating Token Index: 410

Text #3

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.7057. Min Act: -0.1700

Data Index: 552295 (C4 (Web Text))

Max Activating Token Index: 927

Text #4

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6744. Min Act: -0.1700

Data Index: 1641839 (Python Code)

Max Activating Token Index: 1007

Text #5

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.7057. Min Act: -0.1700

Data Index: 1427730 (Python Code)

Max Activating Token Index: 974

Text #6

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6744. Min Act: -0.1700

Data Index: 745397 (C4 (Web Text))

Max Activating Token Index: 656

Text #7

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6118. Min Act: -0.1700

Data Index: 1534795 (Python Code)

Max Activating Token Index: 420

Text #8

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6431. Min Act: -0.1700

Data Index: 857348 (C4 (Web Text))

Max Activating Token Index: 524

Text #9

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6118. Min Act: -0.1700

Data Index: 679697 (C4 (Web Text))

Max Activating Token Index: 330

Text #10

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5805. Min Act: -0.1700

Data Index: 1478776 (Python Code)

Max Activating Token Index: 894

Text #11

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6431. Min Act: -0.1700

Data Index: 1510061 (Python Code)

Max Activating Token Index: 875

Text #12

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5805. Min Act: -0.1700

Data Index: 1265956 (C4 (Web Text))

Max Activating Token Index: 388

Text #13

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6118. Min Act: -0.1700

Data Index: 54280 (C4 (Web Text))

Max Activating Token Index: 192

Text #14

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5491. Min Act: -0.1700

Data Index: 754450 (C4 (Web Text))

Max Activating Token Index: 74

Text #15

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5805. Min Act: -0.1700

Data Index: 1585606 (Python Code)

Max Activating Token Index: 27

Text #16

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5178. Min Act: -0.1700

Data Index: 1461819 (Python Code)

Max Activating Token Index: 786

Text #17

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5805. Min Act: -0.1700

Data Index: 1469100 (Python Code)

Max Activating Token Index: 37

Text #18

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.6118. Min Act: -0.1700

Data Index: 1124369 (C4 (Web Text))

Max Activating Token Index: 173

Text #19

Max Range: 3.7995. Min Range: -3.7995

Max Act: 3.5805. Min Act: -0.1700

Data Index: 888037 (C4 (Web Text))

Max Activating Token Index: 75