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

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

Neuron 1990 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: 5.2626. Min Range: -5.2626

Max Act: 5.2314. Min Act: -0.1700

Data Index: 529733 (C4 (Web Text))

Max Activating Token Index: 662

Text #1

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.2314. Min Act: -0.1700

Data Index: 149042 (C4 (Web Text))

Max Activating Token Index: 484

Text #2

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.2626. Min Act: -0.1700

Data Index: 785065 (C4 (Web Text))

Max Activating Token Index: 560

Text #3

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.2626. Min Act: -0.1700

Data Index: 106202 (C4 (Web Text))

Max Activating Token Index: 70

Text #4

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1689. Min Act: -0.1700

Data Index: 529916 (C4 (Web Text))

Max Activating Token Index: 440

Text #5

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.2001. Min Act: -0.1700

Data Index: 976193 (C4 (Web Text))

Max Activating Token Index: 182

Text #6

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.2001. Min Act: -0.1700

Data Index: 1454653 (Python Code)

Max Activating Token Index: 369

Text #7

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1376. Min Act: -0.1700

Data Index: 1330506 (C4 (Web Text))

Max Activating Token Index: 636

Text #8

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1376. Min Act: -0.1700

Data Index: 311109 (C4 (Web Text))

Max Activating Token Index: 745

Text #9

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1064. Min Act: -0.1700

Data Index: 81961 (C4 (Web Text))

Max Activating Token Index: 944

Text #10

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1064. Min Act: -0.1700

Data Index: 1002549 (C4 (Web Text))

Max Activating Token Index: 430

Text #11

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.0751. Min Act: -0.1700

Data Index: 748321 (C4 (Web Text))

Max Activating Token Index: 93

Text #12

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1376. Min Act: -0.1700

Data Index: 1356813 (C4 (Web Text))

Max Activating Token Index: 842

Text #13

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1064. Min Act: -0.1700

Data Index: 339843 (C4 (Web Text))

Max Activating Token Index: 803

Text #14

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.0751. Min Act: -0.1700

Data Index: 327840 (C4 (Web Text))

Max Activating Token Index: 608

Text #15

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1064. Min Act: -0.1700

Data Index: 1256644 (C4 (Web Text))

Max Activating Token Index: 621

Text #16

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1376. Min Act: -0.1700

Data Index: 1063323 (C4 (Web Text))

Max Activating Token Index: 550

Text #17

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.1376. Min Act: -0.1700

Data Index: 870657 (C4 (Web Text))

Max Activating Token Index: 330

Text #18

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.0751. Min Act: -0.1700

Data Index: 183718 (C4 (Web Text))

Max Activating Token Index: 382

Text #19

Max Range: 5.2626. Min Range: -5.2626

Max Act: 5.0751. Min Act: -0.1700

Data Index: 1132739 (C4 (Web Text))

Max Activating Token Index: 572