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

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

Neuron 429 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.7878. Min Act: -0.1700

Data Index: 447898 (C4 (Web Text))

Max Activating Token Index: 878

Text #1

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.7253. Min Act: -0.1700

Data Index: 1096755 (C4 (Web Text))

Max Activating Token Index: 889

Text #2

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.5690. Min Act: -0.1700

Data Index: 895151 (C4 (Web Text))

Max Activating Token Index: 752

Text #3

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.4753. Min Act: -0.1700

Data Index: 68743 (C4 (Web Text))

Max Activating Token Index: 481

Text #4

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.4753. Min Act: -0.1700

Data Index: 1564635 (Python Code)

Max Activating Token Index: 1023

Text #5

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.4440. Min Act: -0.1700

Data Index: 838638 (C4 (Web Text))

Max Activating Token Index: 541

Text #6

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.5065. Min Act: -0.1700

Data Index: 451053 (C4 (Web Text))

Max Activating Token Index: 741

Text #7

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.4440. Min Act: -0.1700

Data Index: 318937 (C4 (Web Text))

Max Activating Token Index: 548

Text #8

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.2878. Min Act: -0.1700

Data Index: 100411 (C4 (Web Text))

Max Activating Token Index: 502

Text #9

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.3503. Min Act: -0.1700

Data Index: 861309 (C4 (Web Text))

Max Activating Token Index: 584

Text #10

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.2878. Min Act: -0.1700

Data Index: 893972 (C4 (Web Text))

Max Activating Token Index: 703

Text #11

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.2565. Min Act: -0.1700

Data Index: 1480191 (Python Code)

Max Activating Token Index: 370

Text #12

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.1940. Min Act: -0.1700

Data Index: 1475166 (Python Code)

Max Activating Token Index: 314

Text #13

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.0690. Min Act: -0.1700

Data Index: 1588570 (Python Code)

Max Activating Token Index: 1023

Text #14

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.2253. Min Act: -0.1700

Data Index: 454178 (C4 (Web Text))

Max Activating Token Index: 1004

Text #15

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.1940. Min Act: -0.1700

Data Index: 578114 (C4 (Web Text))

Max Activating Token Index: 835

Text #16

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.1628. Min Act: -0.1700

Data Index: 1242969 (C4 (Web Text))

Max Activating Token Index: 176

Text #17

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.1315. Min Act: -0.1700

Data Index: 1526067 (Python Code)

Max Activating Token Index: 466

Text #18

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.1003. Min Act: -0.1700

Data Index: 322951 (C4 (Web Text))

Max Activating Token Index: 314

Text #19

Max Range: 5.7878. Min Range: -5.7878

Max Act: 5.0690. Min Act: -0.1700

Data Index: 1589401 (Python Code)

Max Activating Token Index: 461