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

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

Neuron 1571 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.4859. Min Act: -0.1700

Data Index: 201228 (C4 (Web Text))

Max Activating Token Index: 15

Text #1

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.4859. Min Act: -0.1700

Data Index: 1135933 (C4 (Web Text))

Max Activating Token Index: 962

Text #2

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3918. Min Act: -0.1700

Data Index: 653373 (C4 (Web Text))

Max Activating Token Index: 708

Text #3

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3918. Min Act: -0.1700

Data Index: 494439 (C4 (Web Text))

Max Activating Token Index: 969

Text #4

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3291. Min Act: -0.1700

Data Index: 1322056 (C4 (Web Text))

Max Activating Token Index: 268

Text #5

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3291. Min Act: -0.1700

Data Index: 537980 (C4 (Web Text))

Max Activating Token Index: 375

Text #6

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3133. Min Act: -0.1700

Data Index: 83729 (C4 (Web Text))

Max Activating Token Index: 60

Text #7

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.3447. Min Act: -0.1700

Data Index: 1405746 (Python Code)

Max Activating Token Index: 557

Text #8

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1294765 (C4 (Web Text))

Max Activating Token Index: 364

Text #9

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1275409 (C4 (Web Text))

Max Activating Token Index: 1006

Text #10

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1051757 (C4 (Web Text))

Max Activating Token Index: 510

Text #11

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1621241 (Python Code)

Max Activating Token Index: 335

Text #12

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1426400 (Python Code)

Max Activating Token Index: 378

Text #13

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1041378 (C4 (Web Text))

Max Activating Token Index: 1004

Text #14

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2662. Min Act: -0.1700

Data Index: 1396780 (Python Code)

Max Activating Token Index: 368

Text #15

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2348. Min Act: -0.1700

Data Index: 1538050 (Python Code)

Max Activating Token Index: 513

Text #16

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2662. Min Act: -0.1700

Data Index: 1441337 (Python Code)

Max Activating Token Index: 210

Text #17

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2819. Min Act: -0.1700

Data Index: 1582773 (Python Code)

Max Activating Token Index: 808

Text #18

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2976. Min Act: -0.1700

Data Index: 687356 (C4 (Web Text))

Max Activating Token Index: 896

Text #19

Max Range: 3.4859. Min Range: -3.4859

Max Act: 3.2976. Min Act: -0.1700

Data Index: 685813 (C4 (Web Text))

Max Activating Token Index: 259