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

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

Neuron 81 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: 9.1391. Min Range: -9.1391

Max Act: 9.1391. Min Act: -0.1700

Data Index: 1571523 (Python Code)

Max Activating Token Index: 548

Text #1

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.8891. Min Act: -0.1700

Data Index: 1646639 (Python Code)

Max Activating Token Index: 933

Text #2

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.5766. Min Act: -0.1700

Data Index: 1575766 (Python Code)

Max Activating Token Index: 948

Text #3

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.6391. Min Act: -0.1700

Data Index: 1391678 (Python Code)

Max Activating Token Index: 904

Text #4

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.7016. Min Act: -0.1700

Data Index: 1553730 (Python Code)

Max Activating Token Index: 205

Text #5

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.4516. Min Act: -0.1700

Data Index: 1516455 (Python Code)

Max Activating Token Index: 590

Text #6

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.5141. Min Act: -0.1700

Data Index: 1429335 (Python Code)

Max Activating Token Index: 791

Text #7

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.3266. Min Act: -0.1700

Data Index: 1537216 (Python Code)

Max Activating Token Index: 687

Text #8

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.3266. Min Act: -0.1700

Data Index: 1390472 (Python Code)

Max Activating Token Index: 970

Text #9

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.2641. Min Act: -0.1700

Data Index: 1461784 (Python Code)

Max Activating Token Index: 480

Text #10

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.3266. Min Act: -0.1700

Data Index: 1550023 (Python Code)

Max Activating Token Index: 303

Text #11

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.2016. Min Act: -0.1700

Data Index: 1608619 (Python Code)

Max Activating Token Index: 863

Text #12

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.1391. Min Act: -0.1700

Data Index: 1465831 (Python Code)

Max Activating Token Index: 912

Text #13

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.1391. Min Act: -0.1700

Data Index: 1448030 (Python Code)

Max Activating Token Index: 139

Text #14

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.1391. Min Act: -0.1700

Data Index: 1573228 (Python Code)

Max Activating Token Index: 816

Text #15

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.1391. Min Act: -0.1700

Data Index: 1400484 (Python Code)

Max Activating Token Index: 434

Text #16

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.0766. Min Act: -0.1700

Data Index: 1509804 (Python Code)

Max Activating Token Index: 806

Text #17

Max Range: 9.1391. Min Range: -9.1391

Max Act: 7.9516. Min Act: -0.1700

Data Index: 1505048 (Python Code)

Max Activating Token Index: 696

Text #18

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.0141. Min Act: -0.1700

Data Index: 1533455 (Python Code)

Max Activating Token Index: 153

Text #19

Max Range: 9.1391. Min Range: -9.1391

Max Act: 8.0141. Min Act: -0.1700

Data Index: 1400773 (Python Code)

Max Activating Token Index: 678