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

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

Neuron 1914 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: 5.2729. Min Range: -5.2729

Max Act: 5.2729. Min Act: -0.1700

Data Index: 69407 (C4 (Web Text))

Max Activating Token Index: 302

Text #1

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.2416. Min Act: -0.1700

Data Index: 281728 (C4 (Web Text))

Max Activating Token Index: 118

Text #2

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.1791. Min Act: -0.1700

Data Index: 274843 (C4 (Web Text))

Max Activating Token Index: 848

Text #3

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.1791. Min Act: -0.1700

Data Index: 1268459 (C4 (Web Text))

Max Activating Token Index: 931

Text #4

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.0541. Min Act: -0.1700

Data Index: 38636 (C4 (Web Text))

Max Activating Token Index: 403

Text #5

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.9604. Min Act: -0.1700

Data Index: 90680 (C4 (Web Text))

Max Activating Token Index: 216

Text #6

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.0229. Min Act: -0.1700

Data Index: 885379 (C4 (Web Text))

Max Activating Token Index: 998

Text #7

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.0541. Min Act: -0.1700

Data Index: 61097 (C4 (Web Text))

Max Activating Token Index: 452

Text #8

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.9604. Min Act: -0.1700

Data Index: 533533 (C4 (Web Text))

Max Activating Token Index: 435

Text #9

Max Range: 5.2729. Min Range: -5.2729

Max Act: 5.0541. Min Act: -0.1700

Data Index: 1190343 (C4 (Web Text))

Max Activating Token Index: 157

Text #10

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.9916. Min Act: -0.1700

Data Index: 354314 (C4 (Web Text))

Max Activating Token Index: 417

Text #11

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.9604. Min Act: -0.1700

Data Index: 1486588 (Python Code)

Max Activating Token Index: 980

Text #12

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.9291. Min Act: -0.1700

Data Index: 444282 (C4 (Web Text))

Max Activating Token Index: 856

Text #13

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8979. Min Act: -0.1700

Data Index: 561310 (C4 (Web Text))

Max Activating Token Index: 940

Text #14

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8979. Min Act: -0.1700

Data Index: 572281 (C4 (Web Text))

Max Activating Token Index: 910

Text #15

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8666. Min Act: -0.1700

Data Index: 105255 (C4 (Web Text))

Max Activating Token Index: 966

Text #16

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8979. Min Act: -0.1700

Data Index: 677339 (C4 (Web Text))

Max Activating Token Index: 268

Text #17

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8979. Min Act: -0.1700

Data Index: 279886 (C4 (Web Text))

Max Activating Token Index: 189

Text #18

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8354. Min Act: -0.1700

Data Index: 948240 (C4 (Web Text))

Max Activating Token Index: 909

Text #19

Max Range: 5.2729. Min Range: -5.2729

Max Act: 4.8354. Min Act: -0.1700

Data Index: 596345 (C4 (Web Text))

Max Activating Token Index: 364