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

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

Neuron 1008 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.1732. Min Act: -0.1700

Data Index: 459922 (C4 (Web Text))

Max Activating Token Index: 29

Text #1

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.1102. Min Act: -0.1700

Data Index: 946351 (C4 (Web Text))

Max Activating Token Index: 283

Text #2

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.1102. Min Act: -0.1700

Data Index: 1438095 (Python Code)

Max Activating Token Index: 506

Text #3

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0786. Min Act: -0.1700

Data Index: 1035028 (C4 (Web Text))

Max Activating Token Index: 169

Text #4

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0944. Min Act: -0.1700

Data Index: 1423612 (Python Code)

Max Activating Token Index: 171

Text #5

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0944. Min Act: -0.1700

Data Index: 1413381 (Python Code)

Max Activating Token Index: 116

Text #6

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0944. Min Act: -0.1700

Data Index: 1000822 (C4 (Web Text))

Max Activating Token Index: 261

Text #7

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0786. Min Act: -0.1700

Data Index: 1394731 (Python Code)

Max Activating Token Index: 859

Text #8

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0471. Min Act: -0.1700

Data Index: 1233796 (C4 (Web Text))

Max Activating Token Index: 422

Text #9

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0471. Min Act: -0.1700

Data Index: 1390140 (Python Code)

Max Activating Token Index: 950

Text #10

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 1426285 (Python Code)

Max Activating Token Index: 230

Text #11

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 1465934 (Python Code)

Max Activating Token Index: 410

Text #12

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0471. Min Act: -0.1700

Data Index: 1445121 (Python Code)

Max Activating Token Index: 791

Text #13

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 1578015 (Python Code)

Max Activating Token Index: 765

Text #14

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 309861 (C4 (Web Text))

Max Activating Token Index: 929

Text #15

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 793189 (C4 (Web Text))

Max Activating Token Index: 848

Text #16

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0155. Min Act: -0.1700

Data Index: 475503 (C4 (Web Text))

Max Activating Token Index: 879

Text #17

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 930552 (C4 (Web Text))

Max Activating Token Index: 980

Text #18

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0313. Min Act: -0.1700

Data Index: 940738 (C4 (Web Text))

Max Activating Token Index: 958

Text #19

Max Range: 3.1732. Min Range: -3.1732

Max Act: 3.0471. Min Act: -0.1700

Data Index: 708319 (C4 (Web Text))

Max Activating Token Index: 682