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

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

Neuron 502 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: 7.4915. Min Range: -7.4915

Max Act: 7.4915. Min Act: -0.1700

Data Index: 27207 (C4 (Web Text))

Max Activating Token Index: 751

Text #1

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.3977. Min Act: -0.1700

Data Index: 258126 (C4 (Web Text))

Max Activating Token Index: 543

Text #2

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.2102. Min Act: -0.1700

Data Index: 1510330 (Python Code)

Max Activating Token Index: 460

Text #3

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.2415. Min Act: -0.1700

Data Index: 1194171 (C4 (Web Text))

Max Activating Token Index: 469

Text #4

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.2415. Min Act: -0.1700

Data Index: 1298245 (C4 (Web Text))

Max Activating Token Index: 418

Text #5

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1477. Min Act: -0.1700

Data Index: 1245631 (C4 (Web Text))

Max Activating Token Index: 129

Text #6

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1477. Min Act: -0.1700

Data Index: 477513 (C4 (Web Text))

Max Activating Token Index: 847

Text #7

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1477. Min Act: -0.1700

Data Index: 763212 (C4 (Web Text))

Max Activating Token Index: 215

Text #8

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1165. Min Act: -0.1700

Data Index: 1021690 (C4 (Web Text))

Max Activating Token Index: 350

Text #9

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1165. Min Act: -0.1700

Data Index: 872135 (C4 (Web Text))

Max Activating Token Index: 613

Text #10

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1790. Min Act: -0.1700

Data Index: 1468360 (Python Code)

Max Activating Token Index: 759

Text #11

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1165. Min Act: -0.1700

Data Index: 1374453 (Python Code)

Max Activating Token Index: 678

Text #12

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.1165. Min Act: -0.1700

Data Index: 546022 (C4 (Web Text))

Max Activating Token Index: 919

Text #13

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0852. Min Act: -0.1700

Data Index: 269304 (C4 (Web Text))

Max Activating Token Index: 646

Text #14

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0852. Min Act: -0.1700

Data Index: 317982 (C4 (Web Text))

Max Activating Token Index: 287

Text #15

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0540. Min Act: -0.1700

Data Index: 1026902 (C4 (Web Text))

Max Activating Token Index: 141

Text #16

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0227. Min Act: -0.1700

Data Index: 231870 (C4 (Web Text))

Max Activating Token Index: 332

Text #17

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0852. Min Act: -0.1700

Data Index: 88496 (C4 (Web Text))

Max Activating Token Index: 149

Text #18

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0227. Min Act: -0.1700

Data Index: 1209395 (C4 (Web Text))

Max Activating Token Index: 732

Text #19

Max Range: 7.4915. Min Range: -7.4915

Max Act: 7.0227. Min Act: -0.1700

Data Index: 745101 (C4 (Web Text))

Max Activating Token Index: 345