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

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

Neuron 56 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: 5.5923. Min Range: -5.5923

Max Act: 5.5923. Min Act: -0.1700

Data Index: 1287796 (C4 (Web Text))

Max Activating Token Index: 358

Text #1

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.4361. Min Act: -0.1700

Data Index: 343878 (C4 (Web Text))

Max Activating Token Index: 404

Text #2

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.2486. Min Act: -0.1700

Data Index: 197462 (C4 (Web Text))

Max Activating Token Index: 222

Text #3

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.1861. Min Act: -0.1700

Data Index: 178493 (C4 (Web Text))

Max Activating Token Index: 295

Text #4

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0923. Min Act: -0.1700

Data Index: 606225 (C4 (Web Text))

Max Activating Token Index: 195

Text #5

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.9673. Min Act: -0.1700

Data Index: 1015895 (C4 (Web Text))

Max Activating Token Index: 595

Text #6

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.1236. Min Act: -0.1700

Data Index: 986182 (C4 (Web Text))

Max Activating Token Index: 716

Text #7

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.1548. Min Act: -0.1700

Data Index: 416180 (C4 (Web Text))

Max Activating Token Index: 294

Text #8

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0923. Min Act: -0.1700

Data Index: 747439 (C4 (Web Text))

Max Activating Token Index: 158

Text #9

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0142. Min Act: -0.1700

Data Index: 538086 (C4 (Web Text))

Max Activating Token Index: 931

Text #10

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0298. Min Act: -0.1700

Data Index: 438787 (C4 (Web Text))

Max Activating Token Index: 39

Text #11

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.9986. Min Act: -0.1700

Data Index: 711496 (C4 (Web Text))

Max Activating Token Index: 640

Text #12

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0611. Min Act: -0.1700

Data Index: 648989 (C4 (Web Text))

Max Activating Token Index: 114

Text #13

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0923. Min Act: -0.1700

Data Index: 285038 (C4 (Web Text))

Max Activating Token Index: 289

Text #14

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.9205. Min Act: -0.1700

Data Index: 718736 (C4 (Web Text))

Max Activating Token Index: 209

Text #15

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.9673. Min Act: -0.1700

Data Index: 349569 (C4 (Web Text))

Max Activating Token Index: 135

Text #16

Max Range: 5.5923. Min Range: -5.5923

Max Act: 5.0142. Min Act: -0.1700

Data Index: 253760 (C4 (Web Text))

Max Activating Token Index: 734

Text #17

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.8736. Min Act: -0.1700

Data Index: 394027 (C4 (Web Text))

Max Activating Token Index: 424

Text #18

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.9205. Min Act: -0.1700

Data Index: 1214899 (C4 (Web Text))

Max Activating Token Index: 941

Text #19

Max Range: 5.5923. Min Range: -5.5923

Max Act: 4.8111. Min Act: -0.1700

Data Index: 1208487 (C4 (Web Text))

Max Activating Token Index: 379