Model: GELU Model: 2 Layers, 2048 Neurons per Layer

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

Neuron 506 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-2l')

Text #0

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.8875. Min Act: -0.1700

Data Index: 799696 (C4 (Web Text))

Max Activating Token Index: 431

Text #1

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.8249. Min Act: -0.1700

Data Index: 1579710 (Python Code)

Max Activating Token Index: 783

Text #2

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7937. Min Act: -0.1700

Data Index: 37820 (C4 (Web Text))

Max Activating Token Index: 632

Text #3

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7311. Min Act: -0.1700

Data Index: 1337756 (C4 (Web Text))

Max Activating Token Index: 447

Text #4

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7624. Min Act: -0.1700

Data Index: 1427158 (Python Code)

Max Activating Token Index: 20

Text #5

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7311. Min Act: -0.1700

Data Index: 1548188 (Python Code)

Max Activating Token Index: 613

Text #6

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7624. Min Act: -0.1700

Data Index: 463035 (C4 (Web Text))

Max Activating Token Index: 426

Text #7

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7311. Min Act: -0.1700

Data Index: 248572 (C4 (Web Text))

Max Activating Token Index: 463

Text #8

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6998. Min Act: -0.1700

Data Index: 1495339 (Python Code)

Max Activating Token Index: 26

Text #9

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7624. Min Act: -0.1700

Data Index: 1523063 (Python Code)

Max Activating Token Index: 29

Text #10

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.7311. Min Act: -0.1700

Data Index: 105970 (C4 (Web Text))

Max Activating Token Index: 572

Text #11

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 1445321 (Python Code)

Max Activating Token Index: 407

Text #12

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6372. Min Act: -0.1700

Data Index: 1435199 (Python Code)

Max Activating Token Index: 426

Text #13

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6998. Min Act: -0.1700

Data Index: 1542433 (Python Code)

Max Activating Token Index: 27

Text #14

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 1135628 (C4 (Web Text))

Max Activating Token Index: 754

Text #15

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6372. Min Act: -0.1700

Data Index: 652562 (C4 (Web Text))

Max Activating Token Index: 31

Text #16

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 526089 (C4 (Web Text))

Max Activating Token Index: 677

Text #17

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 500547 (C4 (Web Text))

Max Activating Token Index: 445

Text #18

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 1506374 (Python Code)

Max Activating Token Index: 964

Text #19

Max Range: 3.8875. Min Range: -3.8875

Max Act: 3.6685. Min Act: -0.1700

Data Index: 65844 (C4 (Web Text))

Max Activating Token Index: 315