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

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

Neuron 1548 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.1233. Min Act: -0.1700

Data Index: 10232 (C4 (Web Text))

Max Activating Token Index: 133

Text #1

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.1076. Min Act: -0.1700

Data Index: 1211648 (C4 (Web Text))

Max Activating Token Index: 287

Text #2

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.0445. Min Act: -0.1700

Data Index: 251265 (C4 (Web Text))

Max Activating Token Index: 552

Text #3

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.0287. Min Act: -0.1700

Data Index: 1329193 (C4 (Web Text))

Max Activating Token Index: 691

Text #4

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9971. Min Act: -0.1700

Data Index: 911978 (C4 (Web Text))

Max Activating Token Index: 907

Text #5

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9654. Min Act: -0.1700

Data Index: 131354 (C4 (Web Text))

Max Activating Token Index: 848

Text #6

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9971. Min Act: -0.1700

Data Index: 750275 (C4 (Web Text))

Max Activating Token Index: 577

Text #7

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.0287. Min Act: -0.1700

Data Index: 783744 (C4 (Web Text))

Max Activating Token Index: 700

Text #8

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9971. Min Act: -0.1700

Data Index: 517920 (C4 (Web Text))

Max Activating Token Index: 7

Text #9

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.0129. Min Act: -0.1700

Data Index: 214392 (C4 (Web Text))

Max Activating Token Index: 676

Text #10

Max Range: 3.1233. Min Range: -3.1233

Max Act: 3.0129. Min Act: -0.1700

Data Index: 1016920 (C4 (Web Text))

Max Activating Token Index: 829

Text #11

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9813. Min Act: -0.1700

Data Index: 886635 (C4 (Web Text))

Max Activating Token Index: 485

Text #12

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9813. Min Act: -0.1700

Data Index: 746953 (C4 (Web Text))

Max Activating Token Index: 953

Text #13

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9813. Min Act: -0.1700

Data Index: 1063947 (C4 (Web Text))

Max Activating Token Index: 527

Text #14

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9813. Min Act: -0.1700

Data Index: 1078974 (C4 (Web Text))

Max Activating Token Index: 347

Text #15

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9496. Min Act: -0.1700

Data Index: 632942 (C4 (Web Text))

Max Activating Token Index: 961

Text #16

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9654. Min Act: -0.1700

Data Index: 1255359 (C4 (Web Text))

Max Activating Token Index: 428

Text #17

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9496. Min Act: -0.1700

Data Index: 1318359 (C4 (Web Text))

Max Activating Token Index: 502

Text #18

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9654. Min Act: -0.1700

Data Index: 686172 (C4 (Web Text))

Max Activating Token Index: 351

Text #19

Max Range: 3.1233. Min Range: -3.1233

Max Act: 2.9654. Min Act: -0.1700

Data Index: 35808 (C4 (Web Text))

Max Activating Token Index: 321