Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 14 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 38.1334. Min Range: -38.1334

Max Act: 38.1334. Min Act: -0.0284

Data Index: 937496 (C4 (Web Text))

Max Activating Token Index: 58

Text #1

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.8834. Min Act: -0.0359

Data Index: 427928 (C4 (Web Text))

Max Activating Token Index: 185

Text #2

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.8834. Min Act: -0.0005

Data Index: 863551 (C4 (Web Text))

Max Activating Token Index: 198

Text #3

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.3834. Min Act: -0.0443

Data Index: 1020690 (C4 (Web Text))

Max Activating Token Index: 46

Text #4

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.1334. Min Act: -0.0464

Data Index: 864177 (C4 (Web Text))

Max Activating Token Index: 268

Text #5

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.1334. Min Act: -0.0048

Data Index: 921448 (C4 (Web Text))

Max Activating Token Index: 128

Text #6

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.8834. Min Act: -0.0682

Data Index: 1334676 (C4 (Web Text))

Max Activating Token Index: 286

Text #7

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.8834. Min Act: -0.0042

Data Index: 515889 (C4 (Web Text))

Max Activating Token Index: 285

Text #8

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.8834. Min Act: -0.0807

Data Index: 480367 (C4 (Web Text))

Max Activating Token Index: 124

Text #9

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.6334. Min Act: -0.0031

Data Index: 1013797 (C4 (Web Text))

Max Activating Token Index: 195

Text #10

Max Range: 38.1334. Min Range: -38.1334

Max Act: 37.1334. Min Act: -0.0571

Data Index: 986754 (C4 (Web Text))

Max Activating Token Index: 145

Text #11

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.6334. Min Act: -0.1098

Data Index: 714154 (C4 (Web Text))

Max Activating Token Index: 114

Text #12

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.6334. Min Act: -0.0821

Data Index: 1150119 (C4 (Web Text))

Max Activating Token Index: 299

Text #13

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.3834. Min Act: -0.1612

Data Index: 1073679 (C4 (Web Text))

Max Activating Token Index: 43

Text #14

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.1334. Min Act: -0.0861

Data Index: 395172 (C4 (Web Text))

Max Activating Token Index: 145

Text #15

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.3834. Min Act: -0.1598

Data Index: 996669 (C4 (Web Text))

Max Activating Token Index: 145

Text #16

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.3834. Min Act: -0.0191

Data Index: 749941 (C4 (Web Text))

Max Activating Token Index: 139

Text #17

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.3834. Min Act: -0.0529

Data Index: 233518 (C4 (Web Text))

Max Activating Token Index: 149

Text #18

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.3834. Min Act: -0.0890

Data Index: 580479 (C4 (Web Text))

Max Activating Token Index: 155

Text #19

Max Range: 38.1334. Min Range: -38.1334

Max Act: 36.1334. Min Act: -0.0978

Data Index: 207954 (C4 (Web Text))

Max Activating Token Index: 109