Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 3226 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0490. Min Act: -0.0002

Data Index: 104701 (C4 (Web Text))

Max Activating Token Index: 717

Text #1

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0419. Min Act: -0.0002

Data Index: 822595 (C4 (Web Text))

Max Activating Token Index: 699

Text #2

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0403. Min Act: -0.0002

Data Index: 493625 (C4 (Web Text))

Max Activating Token Index: 998

Text #3

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0411. Min Act: -0.0002

Data Index: 1602284 (Python Code)

Max Activating Token Index: 913

Text #4

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0390. Min Act: -0.0002

Data Index: 414240 (C4 (Web Text))

Max Activating Token Index: 110

Text #5

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0402. Min Act: -0.0002

Data Index: 1025373 (C4 (Web Text))

Max Activating Token Index: 576

Text #6

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0392. Min Act: -0.0002

Data Index: 294041 (C4 (Web Text))

Max Activating Token Index: 665

Text #7

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0387. Min Act: -0.0002

Data Index: 1588236 (Python Code)

Max Activating Token Index: 4

Text #8

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0392. Min Act: -0.0002

Data Index: 897329 (C4 (Web Text))

Max Activating Token Index: 51

Text #9

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0380. Min Act: -0.0002

Data Index: 503105 (C4 (Web Text))

Max Activating Token Index: 241

Text #10

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0390. Min Act: -0.0002

Data Index: 720293 (C4 (Web Text))

Max Activating Token Index: 6

Text #11

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0387. Min Act: -0.0002

Data Index: 1189267 (C4 (Web Text))

Max Activating Token Index: 762

Text #12

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0339. Min Act: -0.0002

Data Index: 727252 (C4 (Web Text))

Max Activating Token Index: 812

Text #13

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0350. Min Act: -0.0002

Data Index: 1146660 (C4 (Web Text))

Max Activating Token Index: 583

Text #14

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0376. Min Act: -0.0002

Data Index: 1234047 (C4 (Web Text))

Max Activating Token Index: 647

Text #15

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0382. Min Act: -0.0002

Data Index: 823148 (C4 (Web Text))

Max Activating Token Index: 162

Text #16

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0381. Min Act: -0.0002

Data Index: 418656 (C4 (Web Text))

Max Activating Token Index: 510

Text #17

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0354. Min Act: -0.0002

Data Index: 1512475 (Python Code)

Max Activating Token Index: 127

Text #18

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0361. Min Act: -0.0002

Data Index: 288598 (C4 (Web Text))

Max Activating Token Index: 168

Text #19

Max Range: 0.0490. Min Range: -0.0490

Max Act: 0.0394. Min Act: -0.0002

Data Index: 93436 (C4 (Web Text))

Max Activating Token Index: 704