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

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

Neuron 1435 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.5535. Min Act: -0.1700

Data Index: 880196 (C4 (Web Text))

Max Activating Token Index: 606

Text #1

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.3660. Min Act: -0.1700

Data Index: 324222 (C4 (Web Text))

Max Activating Token Index: 241

Text #2

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.3347. Min Act: -0.1700

Data Index: 1321003 (C4 (Web Text))

Max Activating Token Index: 208

Text #3

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.3035. Min Act: -0.1700

Data Index: 1382681 (Python Code)

Max Activating Token Index: 893

Text #4

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.3035. Min Act: -0.1700

Data Index: 416745 (C4 (Web Text))

Max Activating Token Index: 954

Text #5

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.2410. Min Act: -0.1700

Data Index: 1392387 (Python Code)

Max Activating Token Index: 460

Text #6

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.2097. Min Act: -0.1700

Data Index: 1625813 (Python Code)

Max Activating Token Index: 182

Text #7

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1785. Min Act: -0.1700

Data Index: 1349627 (C4 (Web Text))

Max Activating Token Index: 388

Text #8

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1785. Min Act: -0.1700

Data Index: 889739 (C4 (Web Text))

Max Activating Token Index: 211

Text #9

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1785. Min Act: -0.1700

Data Index: 264658 (C4 (Web Text))

Max Activating Token Index: 679

Text #10

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1316. Min Act: -0.1700

Data Index: 162974 (C4 (Web Text))

Max Activating Token Index: 535

Text #11

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1160. Min Act: -0.1700

Data Index: 1530185 (Python Code)

Max Activating Token Index: 820

Text #12

Max Range: 7.5535. Min Range: -7.5535

Max Act: 6.9441. Min Act: -0.1700

Data Index: 662077 (C4 (Web Text))

Max Activating Token Index: 415

Text #13

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.1160. Min Act: -0.1700

Data Index: 1542543 (Python Code)

Max Activating Token Index: 283

Text #14

Max Range: 7.5535. Min Range: -7.5535

Max Act: 6.9910. Min Act: -0.1700

Data Index: 222930 (C4 (Web Text))

Max Activating Token Index: 453

Text #15

Max Range: 7.5535. Min Range: -7.5535

Max Act: 6.9441. Min Act: -0.1700

Data Index: 774889 (C4 (Web Text))

Max Activating Token Index: 492

Text #16

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.0066. Min Act: -0.1700

Data Index: 1439016 (Python Code)

Max Activating Token Index: 957

Text #17

Max Range: 7.5535. Min Range: -7.5535

Max Act: 7.0066. Min Act: -0.1700

Data Index: 1411293 (Python Code)

Max Activating Token Index: 538

Text #18

Max Range: 7.5535. Min Range: -7.5535

Max Act: 6.9754. Min Act: -0.1700

Data Index: 1427809 (Python Code)

Max Activating Token Index: 617

Text #19

Max Range: 7.5535. Min Range: -7.5535

Max Act: 6.8191. Min Act: -0.1700

Data Index: 1404808 (Python Code)

Max Activating Token Index: 987