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

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

Neuron 500 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: 5.4802. Min Range: -5.4802

Max Act: 5.3864. Min Act: -0.1700

Data Index: 568842 (C4 (Web Text))

Max Activating Token Index: 320

Text #1

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.4489. Min Act: -0.1700

Data Index: 487425 (C4 (Web Text))

Max Activating Token Index: 680

Text #2

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.4489. Min Act: -0.1700

Data Index: 529346 (C4 (Web Text))

Max Activating Token Index: 351

Text #3

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.4177. Min Act: -0.1700

Data Index: 355203 (C4 (Web Text))

Max Activating Token Index: 74

Text #4

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3864. Min Act: -0.1700

Data Index: 980671 (C4 (Web Text))

Max Activating Token Index: 512

Text #5

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.4177. Min Act: -0.1700

Data Index: 845326 (C4 (Web Text))

Max Activating Token Index: 654

Text #6

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.4802. Min Act: -0.1700

Data Index: 71972 (C4 (Web Text))

Max Activating Token Index: 621

Text #7

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3239. Min Act: -0.1700

Data Index: 722559 (C4 (Web Text))

Max Activating Token Index: 782

Text #8

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3864. Min Act: -0.1700

Data Index: 1257117 (C4 (Web Text))

Max Activating Token Index: 444

Text #9

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3239. Min Act: -0.1700

Data Index: 758937 (C4 (Web Text))

Max Activating Token Index: 390

Text #10

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2927. Min Act: -0.1700

Data Index: 963210 (C4 (Web Text))

Max Activating Token Index: 446

Text #11

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2302. Min Act: -0.1700

Data Index: 91188 (C4 (Web Text))

Max Activating Token Index: 361

Text #12

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3552. Min Act: -0.1700

Data Index: 128190 (C4 (Web Text))

Max Activating Token Index: 518

Text #13

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2302. Min Act: -0.1700

Data Index: 1031482 (C4 (Web Text))

Max Activating Token Index: 394

Text #14

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2927. Min Act: -0.1700

Data Index: 109124 (C4 (Web Text))

Max Activating Token Index: 505

Text #15

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.3239. Min Act: -0.1700

Data Index: 17094 (C4 (Web Text))

Max Activating Token Index: 744

Text #16

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.1677. Min Act: -0.1700

Data Index: 1218297 (C4 (Web Text))

Max Activating Token Index: 365

Text #17

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2927. Min Act: -0.1700

Data Index: 826283 (C4 (Web Text))

Max Activating Token Index: 737

Text #18

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.1677. Min Act: -0.1700

Data Index: 703480 (C4 (Web Text))

Max Activating Token Index: 904

Text #19

Max Range: 5.4802. Min Range: -5.4802

Max Act: 5.2614. Min Act: -0.1700

Data Index: 988546 (C4 (Web Text))

Max Activating Token Index: 588