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

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

Neuron 2001 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: 12.0253. Min Range: -12.0253

Max Act: 12.0253. Min Act: -0.1700

Data Index: 1530590 (Python Code)

Max Activating Token Index: 87

Text #1

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.9003. Min Act: -0.1700

Data Index: 308969 (C4 (Web Text))

Max Activating Token Index: 811

Text #2

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.7753. Min Act: -0.1700

Data Index: 681888 (C4 (Web Text))

Max Activating Token Index: 313

Text #3

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.7753. Min Act: -0.1700

Data Index: 259518 (C4 (Web Text))

Max Activating Token Index: 135

Text #4

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.7753. Min Act: -0.1700

Data Index: 971703 (C4 (Web Text))

Max Activating Token Index: 497

Text #5

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.7128. Min Act: -0.1700

Data Index: 236063 (C4 (Web Text))

Max Activating Token Index: 588

Text #6

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.6503. Min Act: -0.1700

Data Index: 1413185 (Python Code)

Max Activating Token Index: 254

Text #7

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.7128. Min Act: -0.1700

Data Index: 1417313 (Python Code)

Max Activating Token Index: 671

Text #8

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.6503. Min Act: -0.1700

Data Index: 1005618 (C4 (Web Text))

Max Activating Token Index: 61

Text #9

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.6503. Min Act: -0.1699

Data Index: 1087824 (C4 (Web Text))

Max Activating Token Index: 932

Text #10

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.5253. Min Act: -0.1700

Data Index: 1444997 (Python Code)

Max Activating Token Index: 697

Text #11

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.5253. Min Act: -0.1700

Data Index: 915856 (C4 (Web Text))

Max Activating Token Index: 563

Text #12

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.5253. Min Act: -0.1700

Data Index: 1166057 (C4 (Web Text))

Max Activating Token Index: 825

Text #13

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 529886 (C4 (Web Text))

Max Activating Token Index: 146

Text #14

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 1620901 (Python Code)

Max Activating Token Index: 442

Text #15

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 1481741 (Python Code)

Max Activating Token Index: 860

Text #16

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 1553137 (Python Code)

Max Activating Token Index: 446

Text #17

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.5253. Min Act: -0.1700

Data Index: 1041958 (C4 (Web Text))

Max Activating Token Index: 898

Text #18

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 931357 (C4 (Web Text))

Max Activating Token Index: 18

Text #19

Max Range: 12.0253. Min Range: -12.0253

Max Act: 11.4628. Min Act: -0.1700

Data Index: 572605 (C4 (Web Text))

Max Activating Token Index: 623