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

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

Neuron 907 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.4499. Min Act: -0.1700

Data Index: 625736 (C4 (Web Text))

Max Activating Token Index: 495

Text #1

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.3249. Min Act: -0.1699

Data Index: 86995 (C4 (Web Text))

Max Activating Token Index: 384

Text #2

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.2155. Min Act: -0.1700

Data Index: 659657 (C4 (Web Text))

Max Activating Token Index: 484

Text #3

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.2155. Min Act: -0.1700

Data Index: 1351327 (C4 (Web Text))

Max Activating Token Index: 456

Text #4

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1999. Min Act: -0.1700

Data Index: 891738 (C4 (Web Text))

Max Activating Token Index: 912

Text #5

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1687. Min Act: -0.1700

Data Index: 1241388 (C4 (Web Text))

Max Activating Token Index: 651

Text #6

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1374. Min Act: -0.1700

Data Index: 371466 (C4 (Web Text))

Max Activating Token Index: 441

Text #7

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1374. Min Act: -0.1699

Data Index: 1369562 (Python Code)

Max Activating Token Index: 319

Text #8

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1374. Min Act: -0.1700

Data Index: 875964 (C4 (Web Text))

Max Activating Token Index: 405

Text #9

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0905. Min Act: -0.1700

Data Index: 333705 (C4 (Web Text))

Max Activating Token Index: 153

Text #10

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0905. Min Act: -0.1700

Data Index: 256081 (C4 (Web Text))

Max Activating Token Index: 648

Text #11

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1062. Min Act: -0.1700

Data Index: 515626 (C4 (Web Text))

Max Activating Token Index: 193

Text #12

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0749. Min Act: -0.1700

Data Index: 541087 (C4 (Web Text))

Max Activating Token Index: 406

Text #13

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1062. Min Act: -0.1700

Data Index: 532352 (C4 (Web Text))

Max Activating Token Index: 76

Text #14

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1062. Min Act: -0.1700

Data Index: 68446 (C4 (Web Text))

Max Activating Token Index: 123

Text #15

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0593. Min Act: -0.1700

Data Index: 1561962 (Python Code)

Max Activating Token Index: 940

Text #16

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0437. Min Act: -0.1697

Data Index: 769751 (C4 (Web Text))

Max Activating Token Index: 689

Text #17

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0593. Min Act: -0.1700

Data Index: 1233584 (C4 (Web Text))

Max Activating Token Index: 537

Text #18

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.1062. Min Act: -0.1700

Data Index: 1121840 (C4 (Web Text))

Max Activating Token Index: 491

Text #19

Max Range: 5.4499. Min Range: -5.4499

Max Act: 5.0593. Min Act: -0.1700

Data Index: 1384544 (Python Code)

Max Activating Token Index: 859