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

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

Neuron 1640 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: 3.6851. Min Range: -3.6851

Max Act: 3.6538. Min Act: -0.1700

Data Index: 714166 (C4 (Web Text))

Max Activating Token Index: 1012

Text #1

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6538. Min Act: -0.1700

Data Index: 30065 (C4 (Web Text))

Max Activating Token Index: 467

Text #2

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6851. Min Act: -0.1700

Data Index: 548257 (C4 (Web Text))

Max Activating Token Index: 227

Text #3

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 63761 (C4 (Web Text))

Max Activating Token Index: 837

Text #4

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6225. Min Act: -0.1700

Data Index: 43553 (C4 (Web Text))

Max Activating Token Index: 598

Text #5

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6225. Min Act: -0.1700

Data Index: 266529 (C4 (Web Text))

Max Activating Token Index: 527

Text #6

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 1053360 (C4 (Web Text))

Max Activating Token Index: 320

Text #7

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6225. Min Act: -0.1700

Data Index: 1333689 (C4 (Web Text))

Max Activating Token Index: 808

Text #8

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 1023210 (C4 (Web Text))

Max Activating Token Index: 262

Text #9

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 627465 (C4 (Web Text))

Max Activating Token Index: 807

Text #10

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6225. Min Act: -0.1700

Data Index: 1323118 (C4 (Web Text))

Max Activating Token Index: 613

Text #11

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5599. Min Act: -0.1700

Data Index: 1289450 (C4 (Web Text))

Max Activating Token Index: 652

Text #12

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.6225. Min Act: -0.1700

Data Index: 161352 (C4 (Web Text))

Max Activating Token Index: 781

Text #13

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 624561 (C4 (Web Text))

Max Activating Token Index: 452

Text #14

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5286. Min Act: -0.1700

Data Index: 519000 (C4 (Web Text))

Max Activating Token Index: 365

Text #15

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.4972. Min Act: -0.1700

Data Index: 849351 (C4 (Web Text))

Max Activating Token Index: 810

Text #16

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5599. Min Act: -0.1700

Data Index: 736193 (C4 (Web Text))

Max Activating Token Index: 454

Text #17

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5912. Min Act: -0.1700

Data Index: 1200026 (C4 (Web Text))

Max Activating Token Index: 370

Text #18

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.5286. Min Act: -0.1700

Data Index: 680306 (C4 (Web Text))

Max Activating Token Index: 170

Text #19

Max Range: 3.6851. Min Range: -3.6851

Max Act: 3.4659. Min Act: -0.1700

Data Index: 688248 (C4 (Web Text))

Max Activating Token Index: 895