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

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

Neuron 1723 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.7132. Min Act: -0.1700

Data Index: 1037762 (C4 (Web Text))

Max Activating Token Index: 961

Text #1

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.6819. Min Act: -0.1700

Data Index: 895453 (C4 (Web Text))

Max Activating Token Index: 929

Text #2

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.5097. Min Act: -0.1700

Data Index: 289375 (C4 (Web Text))

Max Activating Token Index: 868

Text #3

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.4626. Min Act: -0.1700

Data Index: 475201 (C4 (Web Text))

Max Activating Token Index: 1016

Text #4

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.4783. Min Act: -0.1700

Data Index: 930724 (C4 (Web Text))

Max Activating Token Index: 789

Text #5

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.4626. Min Act: -0.1700

Data Index: 405509 (C4 (Web Text))

Max Activating Token Index: 901

Text #6

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3685. Min Act: -0.1700

Data Index: 1129317 (C4 (Web Text))

Max Activating Token Index: 630

Text #7

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3842. Min Act: -0.1700

Data Index: 1061665 (C4 (Web Text))

Max Activating Token Index: 749

Text #8

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3999. Min Act: -0.1700

Data Index: 27780 (C4 (Web Text))

Max Activating Token Index: 331

Text #9

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3685. Min Act: -0.1700

Data Index: 201342 (C4 (Web Text))

Max Activating Token Index: 502

Text #10

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3685. Min Act: -0.1700

Data Index: 1125094 (C4 (Web Text))

Max Activating Token Index: 990

Text #11

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3842. Min Act: -0.1700

Data Index: 585173 (C4 (Web Text))

Max Activating Token Index: 938

Text #12

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3372. Min Act: -0.1700

Data Index: 647284 (C4 (Web Text))

Max Activating Token Index: 705

Text #13

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3842. Min Act: -0.1700

Data Index: 1158250 (C4 (Web Text))

Max Activating Token Index: 869

Text #14

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3999. Min Act: -0.1700

Data Index: 1074058 (C4 (Web Text))

Max Activating Token Index: 820

Text #15

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3215. Min Act: -0.1700

Data Index: 1363273 (Python Code)

Max Activating Token Index: 652

Text #16

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3685. Min Act: -0.1700

Data Index: 1235725 (C4 (Web Text))

Max Activating Token Index: 810

Text #17

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3685. Min Act: -0.1700

Data Index: 1211264 (C4 (Web Text))

Max Activating Token Index: 668

Text #18

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.4156. Min Act: -0.1700

Data Index: 206605 (C4 (Web Text))

Max Activating Token Index: 594

Text #19

Max Range: 3.7132. Min Range: -3.7132

Max Act: 3.3215. Min Act: -0.1700

Data Index: 806943 (C4 (Web Text))

Max Activating Token Index: 1019