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

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

Neuron 2007 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: 5.8683. Min Range: -5.8683

Max Act: 5.8683. Min Act: -0.1700

Data Index: 406808 (C4 (Web Text))

Max Activating Token Index: 769

Text #1

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.7121. Min Act: -0.1700

Data Index: 892955 (C4 (Web Text))

Max Activating Token Index: 648

Text #2

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.4621. Min Act: -0.1700

Data Index: 483053 (C4 (Web Text))

Max Activating Token Index: 1005

Text #3

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3371. Min Act: -0.1700

Data Index: 341382 (C4 (Web Text))

Max Activating Token Index: 497

Text #4

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3839. Min Act: -0.1700

Data Index: 1204497 (C4 (Web Text))

Max Activating Token Index: 1003

Text #5

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3058. Min Act: -0.1700

Data Index: 763079 (C4 (Web Text))

Max Activating Token Index: 695

Text #6

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3371. Min Act: -0.1700

Data Index: 491559 (C4 (Web Text))

Max Activating Token Index: 795

Text #7

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2902. Min Act: -0.1700

Data Index: 1290098 (C4 (Web Text))

Max Activating Token Index: 756

Text #8

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3371. Min Act: -0.1700

Data Index: 799149 (C4 (Web Text))

Max Activating Token Index: 631

Text #9

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.3683. Min Act: -0.1700

Data Index: 697288 (C4 (Web Text))

Max Activating Token Index: 494

Text #10

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2902. Min Act: -0.1700

Data Index: 100314 (C4 (Web Text))

Max Activating Token Index: 721

Text #11

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2746. Min Act: -0.1700

Data Index: 379141 (C4 (Web Text))

Max Activating Token Index: 719

Text #12

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2746. Min Act: -0.1700

Data Index: 964221 (C4 (Web Text))

Max Activating Token Index: 184

Text #13

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2433. Min Act: -0.1700

Data Index: 801637 (C4 (Web Text))

Max Activating Token Index: 301

Text #14

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2433. Min Act: -0.1700

Data Index: 200262 (C4 (Web Text))

Max Activating Token Index: 508

Text #15

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2902. Min Act: -0.1700

Data Index: 443551 (C4 (Web Text))

Max Activating Token Index: 481

Text #16

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2121. Min Act: -0.1700

Data Index: 279248 (C4 (Web Text))

Max Activating Token Index: 807

Text #17

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2277. Min Act: -0.1700

Data Index: 852140 (C4 (Web Text))

Max Activating Token Index: 209

Text #18

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2746. Min Act: -0.1700

Data Index: 551636 (C4 (Web Text))

Max Activating Token Index: 415

Text #19

Max Range: 5.8683. Min Range: -5.8683

Max Act: 5.2433. Min Act: -0.1700

Data Index: 441897 (C4 (Web Text))

Max Activating Token Index: 278