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

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

Neuron 643 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 7.0275. Min Range: -7.0275

Max Act: 7.0275. Min Act: -0.1700

Data Index: 1147925 (C4 (Web Text))

Max Activating Token Index: 588

Text #1

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.7150. Min Act: -0.1700

Data Index: 1209367 (C4 (Web Text))

Max Activating Token Index: 429

Text #2

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.6837. Min Act: -0.1700

Data Index: 1021394 (C4 (Web Text))

Max Activating Token Index: 369

Text #3

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.5900. Min Act: -0.1700

Data Index: 539211 (C4 (Web Text))

Max Activating Token Index: 636

Text #4

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.6212. Min Act: -0.1700

Data Index: 1352393 (C4 (Web Text))

Max Activating Token Index: 405

Text #5

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.5275. Min Act: -0.1700

Data Index: 330550 (C4 (Web Text))

Max Activating Token Index: 564

Text #6

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.5275. Min Act: -0.1700

Data Index: 638718 (C4 (Web Text))

Max Activating Token Index: 393

Text #7

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.5275. Min Act: -0.1700

Data Index: 205189 (C4 (Web Text))

Max Activating Token Index: 333

Text #8

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.4650. Min Act: -0.1700

Data Index: 398125 (C4 (Web Text))

Max Activating Token Index: 596

Text #9

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.4962. Min Act: -0.1700

Data Index: 151635 (C4 (Web Text))

Max Activating Token Index: 465

Text #10

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.4650. Min Act: -0.1700

Data Index: 1157867 (C4 (Web Text))

Max Activating Token Index: 643

Text #11

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.3712. Min Act: -0.1700

Data Index: 826934 (C4 (Web Text))

Max Activating Token Index: 546

Text #12

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.3400. Min Act: -0.1700

Data Index: 33021 (C4 (Web Text))

Max Activating Token Index: 231

Text #13

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.3712. Min Act: -0.1700

Data Index: 702640 (C4 (Web Text))

Max Activating Token Index: 192

Text #14

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.4025. Min Act: -0.1700

Data Index: 816513 (C4 (Web Text))

Max Activating Token Index: 280

Text #15

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.2462. Min Act: -0.1700

Data Index: 129318 (C4 (Web Text))

Max Activating Token Index: 610

Text #16

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.1837. Min Act: -0.1700

Data Index: 1483583 (Python Code)

Max Activating Token Index: 461

Text #17

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.1837. Min Act: -0.1700

Data Index: 16364 (C4 (Web Text))

Max Activating Token Index: 600

Text #18

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.2462. Min Act: -0.1700

Data Index: 261564 (C4 (Web Text))

Max Activating Token Index: 111

Text #19

Max Range: 7.0275. Min Range: -7.0275

Max Act: 6.2462. Min Act: -0.1700

Data Index: 365367 (C4 (Web Text))

Max Activating Token Index: 130