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

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

Neuron 573 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: 5.5872. Min Range: -5.5872

Max Act: 5.5872. Min Act: -0.1700

Data Index: 96905 (C4 (Web Text))

Max Activating Token Index: 776

Text #1

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.3060. Min Act: -0.1699

Data Index: 1127349 (C4 (Web Text))

Max Activating Token Index: 450

Text #2

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.2435. Min Act: -0.1700

Data Index: 272265 (C4 (Web Text))

Max Activating Token Index: 514

Text #3

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.2747. Min Act: -0.1700

Data Index: 673446 (C4 (Web Text))

Max Activating Token Index: 781

Text #4

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.2435. Min Act: -0.1700

Data Index: 636370 (C4 (Web Text))

Max Activating Token Index: 521

Text #5

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.2122. Min Act: -0.1700

Data Index: 700928 (C4 (Web Text))

Max Activating Token Index: 89

Text #6

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.1497. Min Act: -0.1700

Data Index: 216495 (C4 (Web Text))

Max Activating Token Index: 216

Text #7

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.1810. Min Act: -0.1700

Data Index: 171790 (C4 (Web Text))

Max Activating Token Index: 71

Text #8

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0872. Min Act: -0.1700

Data Index: 1000385 (C4 (Web Text))

Max Activating Token Index: 570

Text #9

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.1497. Min Act: -0.1700

Data Index: 860872 (C4 (Web Text))

Max Activating Token Index: 561

Text #10

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0560. Min Act: -0.1700

Data Index: 1133847 (C4 (Web Text))

Max Activating Token Index: 849

Text #11

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0247. Min Act: -0.1700

Data Index: 947757 (C4 (Web Text))

Max Activating Token Index: 757

Text #12

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0560. Min Act: -0.1700

Data Index: 360641 (C4 (Web Text))

Max Activating Token Index: 941

Text #13

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0247. Min Act: -0.1700

Data Index: 287164 (C4 (Web Text))

Max Activating Token Index: 882

Text #14

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0560. Min Act: -0.1700

Data Index: 1146725 (C4 (Web Text))

Max Activating Token Index: 788

Text #15

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0247. Min Act: -0.1700

Data Index: 836639 (C4 (Web Text))

Max Activating Token Index: 105

Text #16

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0560. Min Act: -0.1700

Data Index: 930639 (C4 (Web Text))

Max Activating Token Index: 836

Text #17

Max Range: 5.5872. Min Range: -5.5872

Max Act: 4.9935. Min Act: -0.1700

Data Index: 1178912 (C4 (Web Text))

Max Activating Token Index: 169

Text #18

Max Range: 5.5872. Min Range: -5.5872

Max Act: 5.0247. Min Act: -0.1700

Data Index: 1190172 (C4 (Web Text))

Max Activating Token Index: 418

Text #19

Max Range: 5.5872. Min Range: -5.5872

Max Act: 4.9935. Min Act: -0.1699

Data Index: 369064 (C4 (Web Text))

Max Activating Token Index: 253