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

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

Neuron 610 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: 5.4781. Min Range: -5.4781

Max Act: 5.4781. Min Act: -0.1700

Data Index: 427110 (C4 (Web Text))

Max Activating Token Index: 498

Text #1

Max Range: 5.4781. Min Range: -5.4781

Max Act: 5.2593. Min Act: -0.1700

Data Index: 893384 (C4 (Web Text))

Max Activating Token Index: 536

Text #2

Max Range: 5.4781. Min Range: -5.4781

Max Act: 5.2906. Min Act: -0.1700

Data Index: 225183 (C4 (Web Text))

Max Activating Token Index: 1023

Text #3

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9468. Min Act: -0.1700

Data Index: 1303065 (C4 (Web Text))

Max Activating Token Index: 465

Text #4

Max Range: 5.4781. Min Range: -5.4781

Max Act: 5.0093. Min Act: -0.1700

Data Index: 113433 (C4 (Web Text))

Max Activating Token Index: 444

Text #5

Max Range: 5.4781. Min Range: -5.4781

Max Act: 5.0406. Min Act: -0.1700

Data Index: 75888 (C4 (Web Text))

Max Activating Token Index: 152

Text #6

Max Range: 5.4781. Min Range: -5.4781

Max Act: 5.2906. Min Act: -0.1700

Data Index: 1061917 (C4 (Web Text))

Max Activating Token Index: 524

Text #7

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.7593. Min Act: -0.1700

Data Index: 937303 (C4 (Web Text))

Max Activating Token Index: 803

Text #8

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9156. Min Act: -0.1700

Data Index: 1317887 (C4 (Web Text))

Max Activating Token Index: 1016

Text #9

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9468. Min Act: -0.1700

Data Index: 995623 (C4 (Web Text))

Max Activating Token Index: 399

Text #10

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9156. Min Act: -0.1700

Data Index: 108351 (C4 (Web Text))

Max Activating Token Index: 712

Text #11

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.6968. Min Act: -0.1700

Data Index: 587936 (C4 (Web Text))

Max Activating Token Index: 488

Text #12

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.8843. Min Act: -0.1700

Data Index: 1353858 (C4 (Web Text))

Max Activating Token Index: 609

Text #13

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9156. Min Act: -0.1700

Data Index: 5170 (C4 (Web Text))

Max Activating Token Index: 566

Text #14

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.8843. Min Act: -0.1700

Data Index: 413440 (C4 (Web Text))

Max Activating Token Index: 315

Text #15

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.7281. Min Act: -0.1700

Data Index: 29373 (C4 (Web Text))

Max Activating Token Index: 521

Text #16

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.8531. Min Act: -0.1700

Data Index: 410629 (C4 (Web Text))

Max Activating Token Index: 402

Text #17

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.9156. Min Act: -0.1700

Data Index: 734668 (C4 (Web Text))

Max Activating Token Index: 667

Text #18

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.7906. Min Act: -0.1700

Data Index: 724760 (C4 (Web Text))

Max Activating Token Index: 644

Text #19

Max Range: 5.4781. Min Range: -5.4781

Max Act: 4.8843. Min Act: -0.1700

Data Index: 726275 (C4 (Web Text))

Max Activating Token Index: 806