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

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

Neuron 1807 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.9021. Min Act: -0.1700

Data Index: 1050949 (C4 (Web Text))

Max Activating Token Index: 629

Text #1

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.9334. Min Act: -0.1700

Data Index: 879406 (C4 (Web Text))

Max Activating Token Index: 1008

Text #2

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.8396. Min Act: -0.1700

Data Index: 7889 (C4 (Web Text))

Max Activating Token Index: 810

Text #3

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.8396. Min Act: -0.1700

Data Index: 389981 (C4 (Web Text))

Max Activating Token Index: 530

Text #4

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.8084. Min Act: -0.1700

Data Index: 1166788 (C4 (Web Text))

Max Activating Token Index: 507

Text #5

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.8396. Min Act: -0.1700

Data Index: 69517 (C4 (Web Text))

Max Activating Token Index: 770

Text #6

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7459. Min Act: -0.1700

Data Index: 574096 (C4 (Web Text))

Max Activating Token Index: 465

Text #7

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.8084. Min Act: -0.1700

Data Index: 1061173 (C4 (Web Text))

Max Activating Token Index: 520

Text #8

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7146. Min Act: -0.1700

Data Index: 48180 (C4 (Web Text))

Max Activating Token Index: 706

Text #9

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7146. Min Act: -0.1700

Data Index: 1615180 (Python Code)

Max Activating Token Index: 592

Text #10

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6834. Min Act: -0.1700

Data Index: 1094740 (C4 (Web Text))

Max Activating Token Index: 460

Text #11

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7459. Min Act: -0.1700

Data Index: 738720 (C4 (Web Text))

Max Activating Token Index: 814

Text #12

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6834. Min Act: -0.1700

Data Index: 1452290 (Python Code)

Max Activating Token Index: 582

Text #13

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7146. Min Act: -0.1700

Data Index: 575558 (C4 (Web Text))

Max Activating Token Index: 296

Text #14

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6521. Min Act: -0.1700

Data Index: 1274388 (C4 (Web Text))

Max Activating Token Index: 326

Text #15

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6834. Min Act: -0.1700

Data Index: 610811 (C4 (Web Text))

Max Activating Token Index: 835

Text #16

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6834. Min Act: -0.1700

Data Index: 885547 (C4 (Web Text))

Max Activating Token Index: 771

Text #17

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7146. Min Act: -0.1700

Data Index: 185677 (C4 (Web Text))

Max Activating Token Index: 790

Text #18

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.6521. Min Act: -0.1700

Data Index: 842945 (C4 (Web Text))

Max Activating Token Index: 748

Text #19

Max Range: 5.9334. Min Range: -5.9334

Max Act: 5.7146. Min Act: -0.1700

Data Index: 1649061 (Python Code)

Max Activating Token Index: 12