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

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

Neuron 160 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: 5.7584. Min Range: -5.7584

Max Act: 5.7584. Min Act: -0.1700

Data Index: 1463466 (Python Code)

Max Activating Token Index: 815

Text #1

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.6334. Min Act: -0.1700

Data Index: 1618084 (Python Code)

Max Activating Token Index: 577

Text #2

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.3521. Min Act: -0.1690

Data Index: 1382304 (Python Code)

Max Activating Token Index: 262

Text #3

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.3209. Min Act: -0.1699

Data Index: 1367920 (Python Code)

Max Activating Token Index: 487

Text #4

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.3209. Min Act: -0.1700

Data Index: 1490385 (Python Code)

Max Activating Token Index: 777

Text #5

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.2584. Min Act: -0.1700

Data Index: 117972 (C4 (Web Text))

Max Activating Token Index: 204

Text #6

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.2271. Min Act: -0.1700

Data Index: 1447972 (Python Code)

Max Activating Token Index: 906

Text #7

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1334. Min Act: -0.1699

Data Index: 1615113 (Python Code)

Max Activating Token Index: 230

Text #8

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1334. Min Act: -0.1698

Data Index: 1465293 (Python Code)

Max Activating Token Index: 146

Text #9

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1334. Min Act: -0.1700

Data Index: 1210010 (C4 (Web Text))

Max Activating Token Index: 777

Text #10

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1646. Min Act: -0.1700

Data Index: 1653080 (Python Code)

Max Activating Token Index: 253

Text #11

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1021. Min Act: -0.1700

Data Index: 1410112 (Python Code)

Max Activating Token Index: 689

Text #12

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1646. Min Act: -0.1700

Data Index: 1426355 (Python Code)

Max Activating Token Index: 351

Text #13

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1646. Min Act: -0.1700

Data Index: 1442731 (Python Code)

Max Activating Token Index: 396

Text #14

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1021. Min Act: -0.1700

Data Index: 1628981 (Python Code)

Max Activating Token Index: 494

Text #15

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1021. Min Act: -0.1700

Data Index: 1596647 (Python Code)

Max Activating Token Index: 290

Text #16

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1021. Min Act: -0.1700

Data Index: 1406892 (Python Code)

Max Activating Token Index: 988

Text #17

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.0709. Min Act: -0.1699

Data Index: 1425093 (Python Code)

Max Activating Token Index: 186

Text #18

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.1021. Min Act: -0.1697

Data Index: 1556147 (Python Code)

Max Activating Token Index: 365

Text #19

Max Range: 5.7584. Min Range: -5.7584

Max Act: 5.0396. Min Act: -0.1700

Data Index: 1529239 (Python Code)

Max Activating Token Index: 636