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

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

Neuron 737 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: 4.1829. Min Range: -4.1829

Max Act: 4.1829. Min Act: -0.1700

Data Index: 1652494 (Python Code)

Max Activating Token Index: 581

Text #1

Max Range: 4.1829. Min Range: -4.1829

Max Act: 4.1360. Min Act: -0.1700

Data Index: 644462 (C4 (Web Text))

Max Activating Token Index: 530

Text #2

Max Range: 4.1829. Min Range: -4.1829

Max Act: 4.1360. Min Act: -0.1700

Data Index: 1409627 (Python Code)

Max Activating Token Index: 809

Text #3

Max Range: 4.1829. Min Range: -4.1829

Max Act: 4.1047. Min Act: -0.1700

Data Index: 307851 (C4 (Web Text))

Max Activating Token Index: 455

Text #4

Max Range: 4.1829. Min Range: -4.1829

Max Act: 4.0109. Min Act: -0.1698

Data Index: 1456846 (Python Code)

Max Activating Token Index: 442

Text #5

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9640. Min Act: -0.1700

Data Index: 1448946 (Python Code)

Max Activating Token Index: 496

Text #6

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9797. Min Act: -0.1700

Data Index: 1537390 (Python Code)

Max Activating Token Index: 881

Text #7

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9484. Min Act: -0.1700

Data Index: 1436332 (Python Code)

Max Activating Token Index: 751

Text #8

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9484. Min Act: -0.1700

Data Index: 448134 (C4 (Web Text))

Max Activating Token Index: 395

Text #9

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9328. Min Act: -0.1700

Data Index: 915602 (C4 (Web Text))

Max Activating Token Index: 864

Text #10

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9328. Min Act: -0.1700

Data Index: 990861 (C4 (Web Text))

Max Activating Token Index: 973

Text #11

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9171. Min Act: -0.1700

Data Index: 1450167 (Python Code)

Max Activating Token Index: 624

Text #12

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9171. Min Act: -0.1700

Data Index: 1435898 (Python Code)

Max Activating Token Index: 628

Text #13

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9640. Min Act: -0.1700

Data Index: 1386950 (Python Code)

Max Activating Token Index: 102

Text #14

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9640. Min Act: -0.1620

Data Index: 1112554 (C4 (Web Text))

Max Activating Token Index: 898

Text #15

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9015. Min Act: -0.1700

Data Index: 1045861 (C4 (Web Text))

Max Activating Token Index: 96

Text #16

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.8702. Min Act: -0.1700

Data Index: 1368751 (Python Code)

Max Activating Token Index: 439

Text #17

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.8859. Min Act: -0.1700

Data Index: 1200719 (C4 (Web Text))

Max Activating Token Index: 573

Text #18

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9328. Min Act: -0.1700

Data Index: 1540894 (Python Code)

Max Activating Token Index: 671

Text #19

Max Range: 4.1829. Min Range: -4.1829

Max Act: 3.9015. Min Act: -0.1700

Data Index: 1504699 (Python Code)

Max Activating Token Index: 447