Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 1708 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0409. Min Act: -0.0001

Data Index: 608104 (C4 (Web Text))

Max Activating Token Index: 642

Text #1

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0378. Min Act: -0.0001

Data Index: 923726 (C4 (Web Text))

Max Activating Token Index: 334

Text #2

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0302. Min Act: -0.0001

Data Index: 492591 (C4 (Web Text))

Max Activating Token Index: 315

Text #3

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0331. Min Act: -0.0001

Data Index: 636144 (C4 (Web Text))

Max Activating Token Index: 794

Text #4

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0334. Min Act: -0.0001

Data Index: 257098 (C4 (Web Text))

Max Activating Token Index: 604

Text #5

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0314. Min Act: -0.0001

Data Index: 785274 (C4 (Web Text))

Max Activating Token Index: 255

Text #6

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0300. Min Act: -0.0001

Data Index: 1579839 (Python Code)

Max Activating Token Index: 669

Text #7

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0295. Min Act: -0.0001

Data Index: 1235826 (C4 (Web Text))

Max Activating Token Index: 215

Text #8

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0298. Min Act: -0.0001

Data Index: 1093733 (C4 (Web Text))

Max Activating Token Index: 400

Text #9

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0298. Min Act: -0.0001

Data Index: 263594 (C4 (Web Text))

Max Activating Token Index: 156

Text #10

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0290. Min Act: -0.0001

Data Index: 978180 (C4 (Web Text))

Max Activating Token Index: 918

Text #11

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0304. Min Act: -0.0001

Data Index: 1365738 (Python Code)

Max Activating Token Index: 1008

Text #12

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0288. Min Act: -0.0001

Data Index: 1101726 (C4 (Web Text))

Max Activating Token Index: 644

Text #13

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0300. Min Act: -0.0001

Data Index: 1418432 (Python Code)

Max Activating Token Index: 505

Text #14

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0274. Min Act: -0.0001

Data Index: 881389 (C4 (Web Text))

Max Activating Token Index: 677

Text #15

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0276. Min Act: -0.0001

Data Index: 1315948 (C4 (Web Text))

Max Activating Token Index: 180

Text #16

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0286. Min Act: -0.0001

Data Index: 335897 (C4 (Web Text))

Max Activating Token Index: 769

Text #17

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0275. Min Act: -0.0001

Data Index: 360365 (C4 (Web Text))

Max Activating Token Index: 344

Text #18

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0268. Min Act: -0.0001

Data Index: 907314 (C4 (Web Text))

Max Activating Token Index: 866

Text #19

Max Range: 0.0409. Min Range: -0.0409

Max Act: 0.0278. Min Act: -0.0001

Data Index: 894272 (C4 (Web Text))

Max Activating Token Index: 466