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

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

Neuron 1635 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.8270. Min Act: -0.1700

Data Index: 1270092 (C4 (Web Text))

Max Activating Token Index: 1

Text #1

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.5923. Min Act: -0.1700

Data Index: 1093277 (C4 (Web Text))

Max Activating Token Index: 2

Text #2

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.4827. Min Act: -0.1699

Data Index: 54698 (C4 (Web Text))

Max Activating Token Index: 1

Text #3

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.4827. Min Act: -0.1700

Data Index: 997613 (C4 (Web Text))

Max Activating Token Index: 1

Text #4

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.4200. Min Act: -0.1700

Data Index: 99794 (C4 (Web Text))

Max Activating Token Index: 1

Text #5

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.4200. Min Act: -0.1700

Data Index: 699296 (C4 (Web Text))

Max Activating Token Index: 1

Text #6

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 1485082 (Python Code)

Max Activating Token Index: 1

Text #7

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 1525895 (Python Code)

Max Activating Token Index: 1

Text #8

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 1584770 (Python Code)

Max Activating Token Index: 1

Text #9

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 1554228 (Python Code)

Max Activating Token Index: 1

Text #10

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 578179 (C4 (Web Text))

Max Activating Token Index: 1

Text #11

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 1140735 (C4 (Web Text))

Max Activating Token Index: 1

Text #12

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1698

Data Index: 264002 (C4 (Web Text))

Max Activating Token Index: 1

Text #13

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 855647 (C4 (Web Text))

Max Activating Token Index: 1

Text #14

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 922973 (C4 (Web Text))

Max Activating Token Index: 1

Text #15

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 808920 (C4 (Web Text))

Max Activating Token Index: 1

Text #16

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 994911 (C4 (Web Text))

Max Activating Token Index: 1

Text #17

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 738490 (C4 (Web Text))

Max Activating Token Index: 1

Text #18

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 86086 (C4 (Web Text))

Max Activating Token Index: 1

Text #19

Max Range: 3.8270. Min Range: -3.8270

Max Act: 3.3415. Min Act: -0.1700

Data Index: 364480 (C4 (Web Text))

Max Activating Token Index: 1