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

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

Neuron 1194 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.1656. Min Range: -3.1656

Max Act: 3.1656. Min Act: -0.1700

Data Index: 1292461 (C4 (Web Text))

Max Activating Token Index: 46

Text #1

Max Range: 3.1656. Min Range: -3.1656

Max Act: 3.1499. Min Act: -0.1700

Data Index: 1061604 (C4 (Web Text))

Max Activating Token Index: 590

Text #2

Max Range: 3.1656. Min Range: -3.1656

Max Act: 3.0237. Min Act: -0.1700

Data Index: 604240 (C4 (Web Text))

Max Activating Token Index: 677

Text #3

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.9762. Min Act: -0.1700

Data Index: 1201651 (C4 (Web Text))

Max Activating Token Index: 353

Text #4

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.9604. Min Act: -0.1700

Data Index: 170760 (C4 (Web Text))

Max Activating Token Index: 484

Text #5

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.8970. Min Act: -0.1700

Data Index: 1432744 (Python Code)

Max Activating Token Index: 1023

Text #6

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.8335. Min Act: -0.1700

Data Index: 1149548 (C4 (Web Text))

Max Activating Token Index: 873

Text #7

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.8335. Min Act: -0.1700

Data Index: 451142 (C4 (Web Text))

Max Activating Token Index: 521

Text #8

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.8176. Min Act: -0.1700

Data Index: 1509291 (Python Code)

Max Activating Token Index: 1023

Text #9

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7857. Min Act: -0.1700

Data Index: 649907 (C4 (Web Text))

Max Activating Token Index: 140

Text #10

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7698. Min Act: -0.1700

Data Index: 356735 (C4 (Web Text))

Max Activating Token Index: 6

Text #11

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7698. Min Act: -0.1700

Data Index: 793035 (C4 (Web Text))

Max Activating Token Index: 169

Text #12

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.8017. Min Act: -0.1700

Data Index: 650284 (C4 (Web Text))

Max Activating Token Index: 292

Text #13

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7857. Min Act: -0.1700

Data Index: 646862 (C4 (Web Text))

Max Activating Token Index: 846

Text #14

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7698. Min Act: -0.1700

Data Index: 770003 (C4 (Web Text))

Max Activating Token Index: 679

Text #15

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7538. Min Act: -0.1700

Data Index: 1266696 (C4 (Web Text))

Max Activating Token Index: 818

Text #16

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7538. Min Act: -0.1700

Data Index: 1109666 (C4 (Web Text))

Max Activating Token Index: 241

Text #17

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7219. Min Act: -0.1700

Data Index: 1108718 (C4 (Web Text))

Max Activating Token Index: 725

Text #18

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7379. Min Act: -0.1700

Data Index: 446251 (C4 (Web Text))

Max Activating Token Index: 538

Text #19

Max Range: 3.1656. Min Range: -3.1656

Max Act: 2.7379. Min Act: -0.1700

Data Index: 1403202 (Python Code)

Max Activating Token Index: 1023