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

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

Neuron 1641 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: 5.3663. Min Range: -5.3663

Max Act: 5.3663. Min Act: -0.1700

Data Index: 697188 (C4 (Web Text))

Max Activating Token Index: 759

Text #1

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.3351. Min Act: -0.1700

Data Index: 1066390 (C4 (Web Text))

Max Activating Token Index: 397

Text #2

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.3038. Min Act: -0.1699

Data Index: 267534 (C4 (Web Text))

Max Activating Token Index: 55

Text #3

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.2726. Min Act: -0.1700

Data Index: 626017 (C4 (Web Text))

Max Activating Token Index: 672

Text #4

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.2413. Min Act: -0.1700

Data Index: 1336506 (C4 (Web Text))

Max Activating Token Index: 685

Text #5

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.2413. Min Act: -0.1699

Data Index: 171793 (C4 (Web Text))

Max Activating Token Index: 19

Text #6

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.2101. Min Act: -0.1700

Data Index: 1036109 (C4 (Web Text))

Max Activating Token Index: 317

Text #7

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.2413. Min Act: -0.1700

Data Index: 409401 (C4 (Web Text))

Max Activating Token Index: 455

Text #8

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1476. Min Act: -0.1700

Data Index: 1184753 (C4 (Web Text))

Max Activating Token Index: 322

Text #9

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1163. Min Act: -0.1700

Data Index: 30970 (C4 (Web Text))

Max Activating Token Index: 739

Text #10

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1788. Min Act: -0.1700

Data Index: 1241141 (C4 (Web Text))

Max Activating Token Index: 561

Text #11

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1163. Min Act: -0.1700

Data Index: 774549 (C4 (Web Text))

Max Activating Token Index: 671

Text #12

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0851. Min Act: -0.1700

Data Index: 1277697 (C4 (Web Text))

Max Activating Token Index: 394

Text #13

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1163. Min Act: -0.1700

Data Index: 73093 (C4 (Web Text))

Max Activating Token Index: 61

Text #14

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0851. Min Act: -0.1700

Data Index: 1192447 (C4 (Web Text))

Max Activating Token Index: 542

Text #15

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0851. Min Act: -0.1700

Data Index: 1623495 (Python Code)

Max Activating Token Index: 196

Text #16

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.1163. Min Act: -0.1700

Data Index: 1180956 (C4 (Web Text))

Max Activating Token Index: 1008

Text #17

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0851. Min Act: -0.1700

Data Index: 1300686 (C4 (Web Text))

Max Activating Token Index: 987

Text #18

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0851. Min Act: -0.1700

Data Index: 853121 (C4 (Web Text))

Max Activating Token Index: 340

Text #19

Max Range: 5.3663. Min Range: -5.3663

Max Act: 5.0538. Min Act: -0.1700

Data Index: 820771 (C4 (Web Text))

Max Activating Token Index: 855