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

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

Neuron 613 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: 3.5441. Min Range: -3.5441

Max Act: 3.5441. Min Act: -0.1700

Data Index: 888573 (C4 (Web Text))

Max Activating Token Index: 684

Text #1

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.5284. Min Act: -0.1700

Data Index: 300122 (C4 (Web Text))

Max Activating Token Index: 933

Text #2

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.4657. Min Act: -0.1700

Data Index: 13337 (C4 (Web Text))

Max Activating Token Index: 300

Text #3

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3559. Min Act: -0.1700

Data Index: 801358 (C4 (Web Text))

Max Activating Token Index: 166

Text #4

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3716. Min Act: -0.1700

Data Index: 1115168 (C4 (Web Text))

Max Activating Token Index: 764

Text #5

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3246. Min Act: -0.1700

Data Index: 1498288 (Python Code)

Max Activating Token Index: 1016

Text #6

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3559. Min Act: -0.1700

Data Index: 1254419 (C4 (Web Text))

Max Activating Token Index: 859

Text #7

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3246. Min Act: -0.1700

Data Index: 1089632 (C4 (Web Text))

Max Activating Token Index: 761

Text #8

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3089. Min Act: -0.1700

Data Index: 108541 (C4 (Web Text))

Max Activating Token Index: 975

Text #9

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.3089. Min Act: -0.1699

Data Index: 1548726 (Python Code)

Max Activating Token Index: 190

Text #10

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2931. Min Act: -0.1700

Data Index: 432507 (C4 (Web Text))

Max Activating Token Index: 424

Text #11

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2617. Min Act: -0.1700

Data Index: 1277834 (C4 (Web Text))

Max Activating Token Index: 985

Text #12

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2774. Min Act: -0.1700

Data Index: 462020 (C4 (Web Text))

Max Activating Token Index: 59

Text #13

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2774. Min Act: -0.1700

Data Index: 1359845 (Python Code)

Max Activating Token Index: 613

Text #14

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2617. Min Act: -0.1700

Data Index: 395210 (C4 (Web Text))

Max Activating Token Index: 933

Text #15

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2460. Min Act: -0.1700

Data Index: 757667 (C4 (Web Text))

Max Activating Token Index: 299

Text #16

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2460. Min Act: -0.1698

Data Index: 1466293 (Python Code)

Max Activating Token Index: 230

Text #17

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2146. Min Act: -0.1700

Data Index: 959209 (C4 (Web Text))

Max Activating Token Index: 829

Text #18

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.2303. Min Act: -0.1700

Data Index: 808605 (C4 (Web Text))

Max Activating Token Index: 999

Text #19

Max Range: 3.5441. Min Range: -3.5441

Max Act: 3.1988. Min Act: -0.1700

Data Index: 714326 (C4 (Web Text))

Max Activating Token Index: 350