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

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

Neuron 132 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.3710. Min Act: -0.1700

Data Index: 1208376 (C4 (Web Text))

Max Activating Token Index: 178

Text #1

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.2925. Min Act: -0.1700

Data Index: 988706 (C4 (Web Text))

Max Activating Token Index: 181

Text #2

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1982. Min Act: -0.1700

Data Index: 654787 (C4 (Web Text))

Max Activating Token Index: 282

Text #3

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1982. Min Act: -0.1700

Data Index: 117431 (C4 (Web Text))

Max Activating Token Index: 991

Text #4

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1352. Min Act: -0.1700

Data Index: 151693 (C4 (Web Text))

Max Activating Token Index: 888

Text #5

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1194. Min Act: -0.1700

Data Index: 987319 (C4 (Web Text))

Max Activating Token Index: 809

Text #6

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1352. Min Act: -0.1700

Data Index: 551608 (C4 (Web Text))

Max Activating Token Index: 717

Text #7

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1194. Min Act: -0.1700

Data Index: 1213278 (C4 (Web Text))

Max Activating Token Index: 597

Text #8

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0879. Min Act: -0.1700

Data Index: 41455 (C4 (Web Text))

Max Activating Token Index: 136

Text #9

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0879. Min Act: -0.1700

Data Index: 959893 (C4 (Web Text))

Max Activating Token Index: 896

Text #10

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0879. Min Act: -0.1700

Data Index: 1071098 (C4 (Web Text))

Max Activating Token Index: 174

Text #11

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.1037. Min Act: -0.1700

Data Index: 265326 (C4 (Web Text))

Max Activating Token Index: 912

Text #12

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0405. Min Act: -0.1700

Data Index: 565022 (C4 (Web Text))

Max Activating Token Index: 894

Text #13

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0405. Min Act: -0.1700

Data Index: 1462918 (Python Code)

Max Activating Token Index: 707

Text #14

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0721. Min Act: -0.1700

Data Index: 1016852 (C4 (Web Text))

Max Activating Token Index: 241

Text #15

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0563. Min Act: -0.1700

Data Index: 39300 (C4 (Web Text))

Max Activating Token Index: 326

Text #16

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0721. Min Act: -0.1700

Data Index: 1045689 (C4 (Web Text))

Max Activating Token Index: 580

Text #17

Max Range: 3.3710. Min Range: -3.3710

Max Act: 2.9773. Min Act: -0.1700

Data Index: 954041 (C4 (Web Text))

Max Activating Token Index: 763

Text #18

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0563. Min Act: -0.1700

Data Index: 315848 (C4 (Web Text))

Max Activating Token Index: 535

Text #19

Max Range: 3.3710. Min Range: -3.3710

Max Act: 3.0405. Min Act: -0.1700

Data Index: 713383 (C4 (Web Text))

Max Activating Token Index: 943