Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 2000 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.9753. Min Act: -0.1700

Data Index: 1181901 (C4 (Web Text))

Max Activating Token Index: 755

Text #1

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.9753. Min Act: -0.1700

Data Index: 1201850 (C4 (Web Text))

Max Activating Token Index: 1023

Text #2

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8502. Min Act: -0.1700

Data Index: 1208241 (C4 (Web Text))

Max Activating Token Index: 1023

Text #3

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8502. Min Act: -0.1700

Data Index: 327084 (C4 (Web Text))

Max Activating Token Index: 691

Text #4

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8815. Min Act: -0.1700

Data Index: 111797 (C4 (Web Text))

Max Activating Token Index: 3

Text #5

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8502. Min Act: -0.1700

Data Index: 389648 (C4 (Web Text))

Max Activating Token Index: 64

Text #6

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 1252137 (C4 (Web Text))

Max Activating Token Index: 715

Text #7

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8502. Min Act: -0.1700

Data Index: 1312778 (C4 (Web Text))

Max Activating Token Index: 1023

Text #8

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 44389 (C4 (Web Text))

Max Activating Token Index: 482

Text #9

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 244029 (C4 (Web Text))

Max Activating Token Index: 688

Text #10

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 509995 (C4 (Web Text))

Max Activating Token Index: 464

Text #11

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7876. Min Act: -0.1700

Data Index: 1017701 (C4 (Web Text))

Max Activating Token Index: 665

Text #12

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7876. Min Act: -0.1700

Data Index: 752574 (C4 (Web Text))

Max Activating Token Index: 1023

Text #13

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7564. Min Act: -0.1700

Data Index: 696274 (C4 (Web Text))

Max Activating Token Index: 247

Text #14

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7564. Min Act: -0.1700

Data Index: 1288466 (C4 (Web Text))

Max Activating Token Index: 694

Text #15

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7876. Min Act: -0.1700

Data Index: 147510 (C4 (Web Text))

Max Activating Token Index: 790

Text #16

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7251. Min Act: -0.1700

Data Index: 348251 (C4 (Web Text))

Max Activating Token Index: 666

Text #17

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 74642 (C4 (Web Text))

Max Activating Token Index: 711

Text #18

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.8189. Min Act: -0.1700

Data Index: 875284 (C4 (Web Text))

Max Activating Token Index: 460

Text #19

Max Range: 3.9753. Min Range: -3.9753

Max Act: 3.7876. Min Act: -0.1700

Data Index: 313506 (C4 (Web Text))

Max Activating Token Index: 389