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

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

Neuron 1600 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: 3.5479. Min Range: -3.5479

Max Act: 3.5479. Min Act: -0.1700

Data Index: 1316883 (C4 (Web Text))

Max Activating Token Index: 440

Text #1

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.5323. Min Act: -0.1700

Data Index: 1239950 (C4 (Web Text))

Max Activating Token Index: 993

Text #2

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4853. Min Act: -0.1700

Data Index: 348489 (C4 (Web Text))

Max Activating Token Index: 653

Text #3

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4382. Min Act: -0.1700

Data Index: 162206 (C4 (Web Text))

Max Activating Token Index: 260

Text #4

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4069. Min Act: -0.1700

Data Index: 275980 (C4 (Web Text))

Max Activating Token Index: 177

Text #5

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4225. Min Act: -0.1700

Data Index: 1231586 (C4 (Web Text))

Max Activating Token Index: 877

Text #6

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4069. Min Act: -0.1700

Data Index: 748219 (C4 (Web Text))

Max Activating Token Index: 999

Text #7

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4225. Min Act: -0.1700

Data Index: 994522 (C4 (Web Text))

Max Activating Token Index: 565

Text #8

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 588548 (C4 (Web Text))

Max Activating Token Index: 820

Text #9

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3755. Min Act: -0.1700

Data Index: 725243 (C4 (Web Text))

Max Activating Token Index: 466

Text #10

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 1227609 (C4 (Web Text))

Max Activating Token Index: 960

Text #11

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4069. Min Act: -0.1700

Data Index: 533310 (C4 (Web Text))

Max Activating Token Index: 717

Text #12

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 1549044 (Python Code)

Max Activating Token Index: 1005

Text #13

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 906193 (C4 (Web Text))

Max Activating Token Index: 588

Text #14

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.4069. Min Act: -0.1700

Data Index: 597161 (C4 (Web Text))

Max Activating Token Index: 894

Text #15

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3755. Min Act: -0.1700

Data Index: 1356299 (C4 (Web Text))

Max Activating Token Index: 985

Text #16

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 1222360 (C4 (Web Text))

Max Activating Token Index: 983

Text #17

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 861410 (C4 (Web Text))

Max Activating Token Index: 1016

Text #18

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 735922 (C4 (Web Text))

Max Activating Token Index: 812

Text #19

Max Range: 3.5479. Min Range: -3.5479

Max Act: 3.3912. Min Act: -0.1700

Data Index: 42317 (C4 (Web Text))

Max Activating Token Index: 312