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

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

Neuron 1575 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.3754. Min Act: -0.1700

Data Index: 597661 (C4 (Web Text))

Max Activating Token Index: 910

Text #1

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.3127. Min Act: -0.1700

Data Index: 497834 (C4 (Web Text))

Max Activating Token Index: 93

Text #2

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.3127. Min Act: -0.1700

Data Index: 341617 (C4 (Web Text))

Max Activating Token Index: 968

Text #3

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2969. Min Act: -0.1700

Data Index: 1549789 (Python Code)

Max Activating Token Index: 276

Text #4

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2969. Min Act: -0.1700

Data Index: 557503 (C4 (Web Text))

Max Activating Token Index: 379

Text #5

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2655. Min Act: -0.1700

Data Index: 146904 (C4 (Web Text))

Max Activating Token Index: 304

Text #6

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2655. Min Act: -0.1700

Data Index: 691291 (C4 (Web Text))

Max Activating Token Index: 280

Text #7

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2341. Min Act: -0.1700

Data Index: 485944 (C4 (Web Text))

Max Activating Token Index: 861

Text #8

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2498. Min Act: -0.1700

Data Index: 1653349 (Python Code)

Max Activating Token Index: 32

Text #9

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2184. Min Act: -0.1700

Data Index: 639485 (C4 (Web Text))

Max Activating Token Index: 518

Text #10

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2026. Min Act: -0.1700

Data Index: 446212 (C4 (Web Text))

Max Activating Token Index: 212

Text #11

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2184. Min Act: -0.1700

Data Index: 1212500 (C4 (Web Text))

Max Activating Token Index: 339

Text #12

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.1869. Min Act: -0.1700

Data Index: 1566359 (Python Code)

Max Activating Token Index: 165

Text #13

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.1869. Min Act: -0.1700

Data Index: 649958 (C4 (Web Text))

Max Activating Token Index: 205

Text #14

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2026. Min Act: -0.1700

Data Index: 375530 (C4 (Web Text))

Max Activating Token Index: 386

Text #15

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.1712. Min Act: -0.1700

Data Index: 969231 (C4 (Web Text))

Max Activating Token Index: 973

Text #16

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2026. Min Act: -0.1700

Data Index: 970653 (C4 (Web Text))

Max Activating Token Index: 107

Text #17

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.1869. Min Act: -0.1700

Data Index: 1157711 (C4 (Web Text))

Max Activating Token Index: 449

Text #18

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2026. Min Act: -0.1700

Data Index: 776001 (C4 (Web Text))

Max Activating Token Index: 1009

Text #19

Max Range: 3.3754. Min Range: -3.3754

Max Act: 3.2184. Min Act: -0.1700

Data Index: 46301 (C4 (Web Text))

Max Activating Token Index: 452