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

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

Neuron 4187 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0399. Min Act: -0.0001

Data Index: 754464 (C4 (Web Text))

Max Activating Token Index: 447

Text #1

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0399. Min Act: -0.0001

Data Index: 30164 (C4 (Web Text))

Max Activating Token Index: 59

Text #2

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0369. Min Act: -0.0001

Data Index: 529196 (C4 (Web Text))

Max Activating Token Index: 722

Text #3

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0245. Min Act: -0.0001

Data Index: 1502562 (Python Code)

Max Activating Token Index: 946

Text #4

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0216. Min Act: -0.0001

Data Index: 1590322 (Python Code)

Max Activating Token Index: 398

Text #5

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0239. Min Act: -0.0001

Data Index: 1343783 (C4 (Web Text))

Max Activating Token Index: 713

Text #6

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0209. Min Act: -0.0001

Data Index: 860570 (C4 (Web Text))

Max Activating Token Index: 1023

Text #7

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0214. Min Act: -0.0001

Data Index: 165857 (C4 (Web Text))

Max Activating Token Index: 706

Text #8

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0234. Min Act: -0.0001

Data Index: 734945 (C4 (Web Text))

Max Activating Token Index: 970

Text #9

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0231. Min Act: -0.0001

Data Index: 1469849 (Python Code)

Max Activating Token Index: 826

Text #10

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0114. Min Act: -0.0001

Data Index: 172125 (C4 (Web Text))

Max Activating Token Index: 232

Text #11

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0216. Min Act: -0.0001

Data Index: 1398436 (Python Code)

Max Activating Token Index: 575

Text #12

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0169. Min Act: -0.0001

Data Index: 465909 (C4 (Web Text))

Max Activating Token Index: 288

Text #13

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0203. Min Act: -0.0001

Data Index: 1100182 (C4 (Web Text))

Max Activating Token Index: 7

Text #14

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0181. Min Act: -0.0001

Data Index: 1367472 (Python Code)

Max Activating Token Index: 830

Text #15

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0155. Min Act: -0.0001

Data Index: 1309492 (C4 (Web Text))

Max Activating Token Index: 344

Text #16

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0164. Min Act: -0.0001

Data Index: 509392 (C4 (Web Text))

Max Activating Token Index: 307

Text #17

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0154. Min Act: -0.0001

Data Index: 338928 (C4 (Web Text))

Max Activating Token Index: 943

Text #18

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0165. Min Act: -0.0001

Data Index: 659083 (C4 (Web Text))

Max Activating Token Index: 139

Text #19

Max Range: 0.0399. Min Range: -0.0399

Max Act: 0.0224. Min Act: -0.0001

Data Index: 313535 (C4 (Web Text))

Max Activating Token Index: 792