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

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

Neuron 40 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: 8.6849. Min Range: -8.6849

Max Act: 8.6849. Min Act: -0.1700

Data Index: 277711 (C4 (Web Text))

Max Activating Token Index: 958

Text #1

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.4974. Min Act: -0.1700

Data Index: 514123 (C4 (Web Text))

Max Activating Token Index: 870

Text #2

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.4036. Min Act: -0.1700

Data Index: 1222639 (C4 (Web Text))

Max Activating Token Index: 457

Text #3

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.4036. Min Act: -0.1700

Data Index: 1574665 (Python Code)

Max Activating Token Index: 536

Text #4

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.3411. Min Act: -0.1700

Data Index: 1275768 (C4 (Web Text))

Max Activating Token Index: 929

Text #5

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.4036. Min Act: -0.1700

Data Index: 669770 (C4 (Web Text))

Max Activating Token Index: 548

Text #6

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.2474. Min Act: -0.1700

Data Index: 1305880 (C4 (Web Text))

Max Activating Token Index: 151

Text #7

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.3724. Min Act: -0.1700

Data Index: 1123054 (C4 (Web Text))

Max Activating Token Index: 457

Text #8

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.2161. Min Act: -0.1700

Data Index: 939281 (C4 (Web Text))

Max Activating Token Index: 774

Text #9

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1536. Min Act: -0.1700

Data Index: 390783 (C4 (Web Text))

Max Activating Token Index: 976

Text #10

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.2474. Min Act: -0.1700

Data Index: 1022495 (C4 (Web Text))

Max Activating Token Index: 380

Text #11

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1224. Min Act: -0.1700

Data Index: 159406 (C4 (Web Text))

Max Activating Token Index: 607

Text #12

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1536. Min Act: -0.1700

Data Index: 1331013 (C4 (Web Text))

Max Activating Token Index: 322

Text #13

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1849. Min Act: -0.1700

Data Index: 266547 (C4 (Web Text))

Max Activating Token Index: 219

Text #14

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1224. Min Act: -0.1700

Data Index: 240091 (C4 (Web Text))

Max Activating Token Index: 87

Text #15

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.1224. Min Act: -0.1700

Data Index: 975544 (C4 (Web Text))

Max Activating Token Index: 682

Text #16

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.0286. Min Act: -0.1700

Data Index: 1255584 (C4 (Web Text))

Max Activating Token Index: 906

Text #17

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.0911. Min Act: -0.1700

Data Index: 254731 (C4 (Web Text))

Max Activating Token Index: 560

Text #18

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.0911. Min Act: -0.1700

Data Index: 1102079 (C4 (Web Text))

Max Activating Token Index: 800

Text #19

Max Range: 8.6849. Min Range: -8.6849

Max Act: 8.0911. Min Act: -0.1700

Data Index: 235645 (C4 (Web Text))

Max Activating Token Index: 307