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

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

Neuron 620 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: 3.7094. Min Range: -3.7094

Max Act: 3.7094. Min Act: -0.1700

Data Index: 1535205 (Python Code)

Max Activating Token Index: 55

Text #1

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 669889 (C4 (Web Text))

Max Activating Token Index: 682

Text #2

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6468. Min Act: -0.1700

Data Index: 14818 (C4 (Web Text))

Max Activating Token Index: 424

Text #3

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6781. Min Act: -0.1700

Data Index: 438727 (C4 (Web Text))

Max Activating Token Index: 913

Text #4

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.7094. Min Act: -0.1700

Data Index: 1459594 (Python Code)

Max Activating Token Index: 627

Text #5

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.5842. Min Act: -0.1700

Data Index: 1586296 (Python Code)

Max Activating Token Index: 94

Text #6

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.5842. Min Act: -0.1700

Data Index: 863530 (C4 (Web Text))

Max Activating Token Index: 794

Text #7

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.7094. Min Act: -0.1700

Data Index: 628982 (C4 (Web Text))

Max Activating Token Index: 429

Text #8

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6468. Min Act: -0.1700

Data Index: 1337069 (C4 (Web Text))

Max Activating Token Index: 945

Text #9

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6781. Min Act: -0.1700

Data Index: 1207293 (C4 (Web Text))

Max Activating Token Index: 791

Text #10

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 46806 (C4 (Web Text))

Max Activating Token Index: 365

Text #11

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 379763 (C4 (Web Text))

Max Activating Token Index: 454

Text #12

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 362642 (C4 (Web Text))

Max Activating Token Index: 930

Text #13

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 806505 (C4 (Web Text))

Max Activating Token Index: 663

Text #14

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.5842. Min Act: -0.1700

Data Index: 1645494 (Python Code)

Max Activating Token Index: 382

Text #15

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6468. Min Act: -0.1700

Data Index: 846306 (C4 (Web Text))

Max Activating Token Index: 771

Text #16

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 813093 (C4 (Web Text))

Max Activating Token Index: 906

Text #17

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 805166 (C4 (Web Text))

Max Activating Token Index: 717

Text #18

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 799536 (C4 (Web Text))

Max Activating Token Index: 511

Text #19

Max Range: 3.7094. Min Range: -3.7094

Max Act: 3.6155. Min Act: -0.1700

Data Index: 45250 (C4 (Web Text))

Max Activating Token Index: 325