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

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

Neuron 527 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: 3.1495. Min Range: -3.1495

Max Act: 3.1338. Min Act: -0.1700

Data Index: 45240 (C4 (Web Text))

Max Activating Token Index: 380

Text #1

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.1495. Min Act: -0.1700

Data Index: 211787 (C4 (Web Text))

Max Activating Token Index: 886

Text #2

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.1495. Min Act: -0.1700

Data Index: 195681 (C4 (Web Text))

Max Activating Token Index: 809

Text #3

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0865. Min Act: -0.1700

Data Index: 976157 (C4 (Web Text))

Max Activating Token Index: 986

Text #4

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.1022. Min Act: -0.1700

Data Index: 744936 (C4 (Web Text))

Max Activating Token Index: 607

Text #5

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0707. Min Act: -0.1700

Data Index: 1596728 (Python Code)

Max Activating Token Index: 589

Text #6

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0865. Min Act: -0.1700

Data Index: 321399 (C4 (Web Text))

Max Activating Token Index: 741

Text #7

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0707. Min Act: -0.1700

Data Index: 1055466 (C4 (Web Text))

Max Activating Token Index: 972

Text #8

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.1180. Min Act: -0.1700

Data Index: 1436790 (Python Code)

Max Activating Token Index: 914

Text #9

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0391. Min Act: -0.1700

Data Index: 573255 (C4 (Web Text))

Max Activating Token Index: 857

Text #10

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0233. Min Act: -0.1700

Data Index: 881468 (C4 (Web Text))

Max Activating Token Index: 886

Text #11

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0391. Min Act: -0.1700

Data Index: 467360 (C4 (Web Text))

Max Activating Token Index: 675

Text #12

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9917. Min Act: -0.1700

Data Index: 92254 (C4 (Web Text))

Max Activating Token Index: 582

Text #13

Max Range: 3.1495. Min Range: -3.1495

Max Act: 3.0075. Min Act: -0.1700

Data Index: 1446214 (Python Code)

Max Activating Token Index: 608

Text #14

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9601. Min Act: -0.1700

Data Index: 1622987 (Python Code)

Max Activating Token Index: 509

Text #15

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9917. Min Act: -0.1700

Data Index: 924791 (C4 (Web Text))

Max Activating Token Index: 840

Text #16

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9601. Min Act: -0.1700

Data Index: 347765 (C4 (Web Text))

Max Activating Token Index: 1020

Text #17

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9759. Min Act: -0.1700

Data Index: 123705 (C4 (Web Text))

Max Activating Token Index: 898

Text #18

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9759. Min Act: -0.1700

Data Index: 324843 (C4 (Web Text))

Max Activating Token Index: 817

Text #19

Max Range: 3.1495. Min Range: -3.1495

Max Act: 2.9442. Min Act: -0.1700

Data Index: 140513 (C4 (Web Text))

Max Activating Token Index: 411