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

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

Neuron 563 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: 4.0320. Min Range: -4.0320

Max Act: 4.0320. Min Act: -0.1700

Data Index: 1376414 (Python Code)

Max Activating Token Index: 729

Text #1

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.9694. Min Act: -0.1700

Data Index: 1311467 (C4 (Web Text))

Max Activating Token Index: 6

Text #2

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.8756. Min Act: -0.1700

Data Index: 286269 (C4 (Web Text))

Max Activating Token Index: 461

Text #3

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.8444. Min Act: -0.1699

Data Index: 531269 (C4 (Web Text))

Max Activating Token Index: 225

Text #4

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7818. Min Act: -0.1700

Data Index: 1002023 (C4 (Web Text))

Max Activating Token Index: 946

Text #5

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7818. Min Act: -0.1699

Data Index: 734740 (C4 (Web Text))

Max Activating Token Index: 233

Text #6

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7192. Min Act: -0.1700

Data Index: 1366792 (Python Code)

Max Activating Token Index: 623

Text #7

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7192. Min Act: -0.1700

Data Index: 1490314 (Python Code)

Max Activating Token Index: 226

Text #8

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7505. Min Act: -0.1699

Data Index: 656961 (C4 (Web Text))

Max Activating Token Index: 44

Text #9

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7505. Min Act: -0.1700

Data Index: 1167104 (C4 (Web Text))

Max Activating Token Index: 391

Text #10

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7505. Min Act: -0.1700

Data Index: 1018846 (C4 (Web Text))

Max Activating Token Index: 65

Text #11

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7192. Min Act: -0.1700

Data Index: 1449770 (Python Code)

Max Activating Token Index: 39

Text #12

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7192. Min Act: -0.1696

Data Index: 743370 (C4 (Web Text))

Max Activating Token Index: 61

Text #13

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6879. Min Act: -0.1700

Data Index: 1140445 (C4 (Web Text))

Max Activating Token Index: 679

Text #14

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.7192. Min Act: -0.1699

Data Index: 969226 (C4 (Web Text))

Max Activating Token Index: 104

Text #15

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6879. Min Act: -0.1699

Data Index: 664824 (C4 (Web Text))

Max Activating Token Index: 496

Text #16

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6566. Min Act: -0.1700

Data Index: 798995 (C4 (Web Text))

Max Activating Token Index: 909

Text #17

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6566. Min Act: -0.1700

Data Index: 860679 (C4 (Web Text))

Max Activating Token Index: 993

Text #18

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6566. Min Act: -0.1699

Data Index: 1420087 (Python Code)

Max Activating Token Index: 909

Text #19

Max Range: 4.0320. Min Range: -4.0320

Max Act: 3.6566. Min Act: -0.1700

Data Index: 1426032 (Python Code)

Max Activating Token Index: 528