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

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

Neuron 1008 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: 3.6317. Min Range: -3.6317

Max Act: 3.6317. Min Act: -0.1700

Data Index: 358062 (C4 (Web Text))

Max Activating Token Index: 126

Text #1

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.6160. Min Act: -0.1700

Data Index: 1000842 (C4 (Web Text))

Max Activating Token Index: 981

Text #2

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5690. Min Act: -0.1700

Data Index: 1393301 (Python Code)

Max Activating Token Index: 9

Text #3

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5220. Min Act: -0.1700

Data Index: 1250731 (C4 (Web Text))

Max Activating Token Index: 476

Text #4

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5064. Min Act: -0.1700

Data Index: 212162 (C4 (Web Text))

Max Activating Token Index: 47

Text #5

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5064. Min Act: -0.1700

Data Index: 1302196 (C4 (Web Text))

Max Activating Token Index: 512

Text #6

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5064. Min Act: -0.1700

Data Index: 1586490 (Python Code)

Max Activating Token Index: 172

Text #7

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.5220. Min Act: -0.1700

Data Index: 42051 (C4 (Web Text))

Max Activating Token Index: 1012

Text #8

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4750. Min Act: -0.1700

Data Index: 1521954 (Python Code)

Max Activating Token Index: 23

Text #9

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4907. Min Act: -0.1700

Data Index: 710868 (C4 (Web Text))

Max Activating Token Index: 371

Text #10

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4437. Min Act: -0.1700

Data Index: 1098310 (C4 (Web Text))

Max Activating Token Index: 374

Text #11

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4750. Min Act: -0.1700

Data Index: 312931 (C4 (Web Text))

Max Activating Token Index: 439

Text #12

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4594. Min Act: -0.1700

Data Index: 1640350 (Python Code)

Max Activating Token Index: 898

Text #13

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4594. Min Act: -0.1700

Data Index: 834426 (C4 (Web Text))

Max Activating Token Index: 480

Text #14

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4750. Min Act: -0.1700

Data Index: 398601 (C4 (Web Text))

Max Activating Token Index: 312

Text #15

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4750. Min Act: -0.1700

Data Index: 1359436 (C4 (Web Text))

Max Activating Token Index: 950

Text #16

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4594. Min Act: -0.1700

Data Index: 373004 (C4 (Web Text))

Max Activating Token Index: 317

Text #17

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4437. Min Act: -0.1700

Data Index: 1259787 (C4 (Web Text))

Max Activating Token Index: 866

Text #18

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4437. Min Act: -0.1700

Data Index: 604530 (C4 (Web Text))

Max Activating Token Index: 467

Text #19

Max Range: 3.6317. Min Range: -3.6317

Max Act: 3.4437. Min Act: -0.1700

Data Index: 1589437 (Python Code)

Max Activating Token Index: 350