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

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

Neuron 2013 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: 5.4392. Min Range: -5.4392

Max Act: 5.4392. Min Act: -0.1700

Data Index: 511013 (C4 (Web Text))

Max Activating Token Index: 50

Text #1

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.4392. Min Act: -0.1700

Data Index: 64549 (C4 (Web Text))

Max Activating Token Index: 700

Text #2

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.3142. Min Act: -0.1700

Data Index: 1274091 (C4 (Web Text))

Max Activating Token Index: 951

Text #3

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.2829. Min Act: -0.1700

Data Index: 698125 (C4 (Web Text))

Max Activating Token Index: 456

Text #4

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.2204. Min Act: -0.1700

Data Index: 1111480 (C4 (Web Text))

Max Activating Token Index: 385

Text #5

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1579. Min Act: -0.1700

Data Index: 635820 (C4 (Web Text))

Max Activating Token Index: 718

Text #6

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1892. Min Act: -0.1700

Data Index: 1201621 (C4 (Web Text))

Max Activating Token Index: 678

Text #7

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1892. Min Act: -0.1700

Data Index: 550600 (C4 (Web Text))

Max Activating Token Index: 371

Text #8

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1579. Min Act: -0.1700

Data Index: 680519 (C4 (Web Text))

Max Activating Token Index: 963

Text #9

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1267. Min Act: -0.1700

Data Index: 298150 (C4 (Web Text))

Max Activating Token Index: 470

Text #10

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1267. Min Act: -0.1700

Data Index: 103413 (C4 (Web Text))

Max Activating Token Index: 404

Text #11

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0954. Min Act: -0.1700

Data Index: 1199687 (C4 (Web Text))

Max Activating Token Index: 97

Text #12

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1267. Min Act: -0.1700

Data Index: 224220 (C4 (Web Text))

Max Activating Token Index: 194

Text #13

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.1267. Min Act: -0.1700

Data Index: 289251 (C4 (Web Text))

Max Activating Token Index: 487

Text #14

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0329. Min Act: -0.1700

Data Index: 100973 (C4 (Web Text))

Max Activating Token Index: 102

Text #15

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0642. Min Act: -0.1700

Data Index: 241947 (C4 (Web Text))

Max Activating Token Index: 338

Text #16

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0642. Min Act: -0.1700

Data Index: 461361 (C4 (Web Text))

Max Activating Token Index: 189

Text #17

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0642. Min Act: -0.1700

Data Index: 554489 (C4 (Web Text))

Max Activating Token Index: 568

Text #18

Max Range: 5.4392. Min Range: -5.4392

Max Act: 5.0017. Min Act: -0.1700

Data Index: 39579 (C4 (Web Text))

Max Activating Token Index: 866

Text #19

Max Range: 5.4392. Min Range: -5.4392

Max Act: 4.9704. Min Act: -0.1700

Data Index: 310396 (C4 (Web Text))

Max Activating Token Index: 513