Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 576 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 1.1623. Min Range: -1.1623

Max Act: 1.1623. Min Act: -0.0002

Data Index: 426015 (C4 (Web Text))

Max Activating Token Index: 940

Text #1

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.8855. Min Act: -0.0002

Data Index: 1520298 (Python Code)

Max Activating Token Index: 887

Text #2

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.8503. Min Act: -0.0002

Data Index: 1388406 (Python Code)

Max Activating Token Index: 960

Text #3

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.6474. Min Act: -0.0002

Data Index: 471528 (C4 (Web Text))

Max Activating Token Index: 555

Text #4

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.6155. Min Act: -0.0002

Data Index: 813099 (C4 (Web Text))

Max Activating Token Index: 573

Text #5

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.5257. Min Act: -0.0002

Data Index: 111582 (C4 (Web Text))

Max Activating Token Index: 219

Text #6

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.5611. Min Act: -0.0002

Data Index: 917444 (C4 (Web Text))

Max Activating Token Index: 251

Text #7

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.5384. Min Act: -0.0002

Data Index: 1378617 (Python Code)

Max Activating Token Index: 481

Text #8

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.5363. Min Act: -0.0002

Data Index: 400947 (C4 (Web Text))

Max Activating Token Index: 707

Text #9

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4949. Min Act: -0.0002

Data Index: 286578 (C4 (Web Text))

Max Activating Token Index: 480

Text #10

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4779. Min Act: -0.0002

Data Index: 1380808 (Python Code)

Max Activating Token Index: 1023

Text #11

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4508. Min Act: -0.0002

Data Index: 969370 (C4 (Web Text))

Max Activating Token Index: 256

Text #12

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4718. Min Act: -0.0002

Data Index: 818911 (C4 (Web Text))

Max Activating Token Index: 238

Text #13

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4306. Min Act: -0.0002

Data Index: 1451901 (Python Code)

Max Activating Token Index: 921

Text #14

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4450. Min Act: -0.0002

Data Index: 1012377 (C4 (Web Text))

Max Activating Token Index: 288

Text #15

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4144. Min Act: -0.0002

Data Index: 1428510 (Python Code)

Max Activating Token Index: 836

Text #16

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4200. Min Act: -0.0002

Data Index: 1515840 (Python Code)

Max Activating Token Index: 581

Text #17

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4330. Min Act: -0.0002

Data Index: 218343 (C4 (Web Text))

Max Activating Token Index: 543

Text #18

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4294. Min Act: -0.0002

Data Index: 1402815 (Python Code)

Max Activating Token Index: 833

Text #19

Max Range: 1.1623. Min Range: -1.1623

Max Act: 0.4093. Min Act: -0.0002

Data Index: 1394951 (Python Code)

Max Activating Token Index: 644