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

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

Neuron 1274 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: 5.3170. Min Range: -5.3170

Max Act: 5.3170. Min Act: -0.1700

Data Index: 321434 (C4 (Web Text))

Max Activating Token Index: 781

Text #1

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2857. Min Act: -0.1700

Data Index: 625809 (C4 (Web Text))

Max Activating Token Index: 280

Text #2

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2232. Min Act: -0.1700

Data Index: 18088 (C4 (Web Text))

Max Activating Token Index: 234

Text #3

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2232. Min Act: -0.1700

Data Index: 555630 (C4 (Web Text))

Max Activating Token Index: 857

Text #4

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2232. Min Act: -0.1699

Data Index: 286691 (C4 (Web Text))

Max Activating Token Index: 386

Text #5

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2232. Min Act: -0.1699

Data Index: 1469622 (Python Code)

Max Activating Token Index: 605

Text #6

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 949553 (C4 (Web Text))

Max Activating Token Index: 675

Text #7

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.2545. Min Act: -0.1700

Data Index: 1165427 (C4 (Web Text))

Max Activating Token Index: 1004

Text #8

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 97239 (C4 (Web Text))

Max Activating Token Index: 980

Text #9

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1607. Min Act: -0.1700

Data Index: 70744 (C4 (Web Text))

Max Activating Token Index: 207

Text #10

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1607. Min Act: -0.1700

Data Index: 553140 (C4 (Web Text))

Max Activating Token Index: 287

Text #11

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1607. Min Act: -0.1700

Data Index: 733482 (C4 (Web Text))

Max Activating Token Index: 732

Text #12

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 611111 (C4 (Web Text))

Max Activating Token Index: 785

Text #13

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1295. Min Act: -0.1700

Data Index: 1205823 (C4 (Web Text))

Max Activating Token Index: 493

Text #14

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1295. Min Act: -0.1700

Data Index: 151962 (C4 (Web Text))

Max Activating Token Index: 545

Text #15

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 208826 (C4 (Web Text))

Max Activating Token Index: 341

Text #16

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 1243011 (C4 (Web Text))

Max Activating Token Index: 112

Text #17

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1920. Min Act: -0.1700

Data Index: 672915 (C4 (Web Text))

Max Activating Token Index: 870

Text #18

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1607. Min Act: -0.1700

Data Index: 822163 (C4 (Web Text))

Max Activating Token Index: 657

Text #19

Max Range: 5.3170. Min Range: -5.3170

Max Act: 5.1607. Min Act: -0.1700

Data Index: 1435970 (Python Code)

Max Activating Token Index: 338