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

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

Neuron 12 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: 3.3934. Min Range: -3.3934

Max Act: 3.3934. Min Act: -0.1700

Data Index: 572229 (C4 (Web Text))

Max Activating Token Index: 924

Text #1

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.2992. Min Act: -0.1700

Data Index: 15966 (C4 (Web Text))

Max Activating Token Index: 858

Text #2

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.2364. Min Act: -0.1700

Data Index: 68624 (C4 (Web Text))

Max Activating Token Index: 638

Text #3

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.2364. Min Act: -0.1700

Data Index: 1021672 (C4 (Web Text))

Max Activating Token Index: 973

Text #4

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.2364. Min Act: -0.1700

Data Index: 1192663 (C4 (Web Text))

Max Activating Token Index: 34

Text #5

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.1892. Min Act: -0.1700

Data Index: 658221 (C4 (Web Text))

Max Activating Token Index: 661

Text #6

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.1577. Min Act: -0.1700

Data Index: 105899 (C4 (Web Text))

Max Activating Token Index: 481

Text #7

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.1262. Min Act: -0.1700

Data Index: 806611 (C4 (Web Text))

Max Activating Token Index: 92

Text #8

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.1262. Min Act: -0.1700

Data Index: 544491 (C4 (Web Text))

Max Activating Token Index: 777

Text #9

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0789. Min Act: -0.1700

Data Index: 1573156 (Python Code)

Max Activating Token Index: 581

Text #10

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.1104. Min Act: -0.1700

Data Index: 952270 (C4 (Web Text))

Max Activating Token Index: 506

Text #11

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0947. Min Act: -0.1700

Data Index: 719780 (C4 (Web Text))

Max Activating Token Index: 898

Text #12

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0789. Min Act: -0.1700

Data Index: 305187 (C4 (Web Text))

Max Activating Token Index: 907

Text #13

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0631. Min Act: -0.1700

Data Index: 959646 (C4 (Web Text))

Max Activating Token Index: 653

Text #14

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0315. Min Act: -0.1700

Data Index: 822452 (C4 (Web Text))

Max Activating Token Index: 771

Text #15

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0631. Min Act: -0.1700

Data Index: 86959 (C4 (Web Text))

Max Activating Token Index: 94

Text #16

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0473. Min Act: -0.1700

Data Index: 8562 (C4 (Web Text))

Max Activating Token Index: 750

Text #17

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0631. Min Act: -0.1700

Data Index: 40856 (C4 (Web Text))

Max Activating Token Index: 800

Text #18

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0157. Min Act: -0.1700

Data Index: 28685 (C4 (Web Text))

Max Activating Token Index: 599

Text #19

Max Range: 3.3934. Min Range: -3.3934

Max Act: 3.0315. Min Act: -0.1700

Data Index: 639535 (C4 (Web Text))

Max Activating Token Index: 573