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

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

Neuron 1500 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: 3.8414. Min Range: -3.8414

Max Act: 3.8414. Min Act: -0.1700

Data Index: 796688 (C4 (Web Text))

Max Activating Token Index: 667

Text #1

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.8414. Min Act: -0.1700

Data Index: 917013 (C4 (Web Text))

Max Activating Token Index: 49

Text #2

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.8414. Min Act: -0.1700

Data Index: 1163448 (C4 (Web Text))

Max Activating Token Index: 1018

Text #3

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7789. Min Act: -0.1700

Data Index: 264975 (C4 (Web Text))

Max Activating Token Index: 923

Text #4

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7789. Min Act: -0.1700

Data Index: 1163337 (C4 (Web Text))

Max Activating Token Index: 603

Text #5

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7320. Min Act: -0.1700

Data Index: 965906 (C4 (Web Text))

Max Activating Token Index: 714

Text #6

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7632. Min Act: -0.1700

Data Index: 367684 (C4 (Web Text))

Max Activating Token Index: 762

Text #7

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7320. Min Act: -0.1700

Data Index: 603617 (C4 (Web Text))

Max Activating Token Index: 160

Text #8

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7476. Min Act: -0.1700

Data Index: 721603 (C4 (Web Text))

Max Activating Token Index: 500

Text #9

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7007. Min Act: -0.1700

Data Index: 1050356 (C4 (Web Text))

Max Activating Token Index: 618

Text #10

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7007. Min Act: -0.1700

Data Index: 28881 (C4 (Web Text))

Max Activating Token Index: 751

Text #11

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.7163. Min Act: -0.1700

Data Index: 1196143 (C4 (Web Text))

Max Activating Token Index: 982

Text #12

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6850. Min Act: -0.1700

Data Index: 1049938 (C4 (Web Text))

Max Activating Token Index: 110

Text #13

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6694. Min Act: -0.1700

Data Index: 126785 (C4 (Web Text))

Max Activating Token Index: 831

Text #14

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6694. Min Act: -0.1700

Data Index: 1223151 (C4 (Web Text))

Max Activating Token Index: 881

Text #15

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6067. Min Act: -0.1700

Data Index: 805360 (C4 (Web Text))

Max Activating Token Index: 426

Text #16

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6067. Min Act: -0.1700

Data Index: 140153 (C4 (Web Text))

Max Activating Token Index: 941

Text #17

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.5911. Min Act: -0.1700

Data Index: 545721 (C4 (Web Text))

Max Activating Token Index: 15

Text #18

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6224. Min Act: -0.1700

Data Index: 1128925 (C4 (Web Text))

Max Activating Token Index: 651

Text #19

Max Range: 3.8414. Min Range: -3.8414

Max Act: 3.6224. Min Act: -0.1700

Data Index: 891018 (C4 (Web Text))

Max Activating Token Index: 687