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

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

Neuron 2021 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.4832. Min Act: -0.1700

Data Index: 1143078 (C4 (Web Text))

Max Activating Token Index: 254

Text #1

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.1692. Min Act: -0.1700

Data Index: 381270 (C4 (Web Text))

Max Activating Token Index: 340

Text #2

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.1377. Min Act: -0.1700

Data Index: 3527 (C4 (Web Text))

Max Activating Token Index: 240

Text #3

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.1062. Min Act: -0.1700

Data Index: 42789 (C4 (Web Text))

Max Activating Token Index: 868

Text #4

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0746. Min Act: -0.1700

Data Index: 565126 (C4 (Web Text))

Max Activating Token Index: 736

Text #5

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0746. Min Act: -0.1700

Data Index: 1211884 (C4 (Web Text))

Max Activating Token Index: 318

Text #6

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0589. Min Act: -0.1700

Data Index: 1313429 (C4 (Web Text))

Max Activating Token Index: 386

Text #7

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0904. Min Act: -0.1700

Data Index: 315295 (C4 (Web Text))

Max Activating Token Index: 923

Text #8

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0589. Min Act: -0.1700

Data Index: 915798 (C4 (Web Text))

Max Activating Token Index: 175

Text #9

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0589. Min Act: -0.1700

Data Index: 1177446 (C4 (Web Text))

Max Activating Token Index: 181

Text #10

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0273. Min Act: -0.1700

Data Index: 175098 (C4 (Web Text))

Max Activating Token Index: 760

Text #11

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0904. Min Act: -0.1700

Data Index: 877602 (C4 (Web Text))

Max Activating Token Index: 817

Text #12

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0115. Min Act: -0.1700

Data Index: 784461 (C4 (Web Text))

Max Activating Token Index: 941

Text #13

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0431. Min Act: -0.1700

Data Index: 436219 (C4 (Web Text))

Max Activating Token Index: 716

Text #14

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0431. Min Act: -0.1700

Data Index: 710653 (C4 (Web Text))

Max Activating Token Index: 401

Text #15

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0273. Min Act: -0.1700

Data Index: 1017292 (C4 (Web Text))

Max Activating Token Index: 266

Text #16

Max Range: 3.4832. Min Range: -3.4832

Max Act: 2.9957. Min Act: -0.1700

Data Index: 1061475 (C4 (Web Text))

Max Activating Token Index: 842

Text #17

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0273. Min Act: -0.1700

Data Index: 151503 (C4 (Web Text))

Max Activating Token Index: 496

Text #18

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0431. Min Act: -0.1700

Data Index: 476521 (C4 (Web Text))

Max Activating Token Index: 1000

Text #19

Max Range: 3.4832. Min Range: -3.4832

Max Act: 3.0746. Min Act: -0.1700

Data Index: 416735 (C4 (Web Text))

Max Activating Token Index: 364