Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 2845 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0295. Min Act: -0.0003

Data Index: 280054 (C4 (Web Text))

Max Activating Token Index: 328

Text #1

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0175. Min Act: -0.0003

Data Index: 34710 (C4 (Web Text))

Max Activating Token Index: 398

Text #2

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0218. Min Act: -0.0003

Data Index: 990550 (C4 (Web Text))

Max Activating Token Index: 18

Text #3

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0357. Min Act: -0.0003

Data Index: 1581359 (Python Code)

Max Activating Token Index: 1023

Text #4

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0289. Min Act: -0.0003

Data Index: 45430 (C4 (Web Text))

Max Activating Token Index: 359

Text #5

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0291. Min Act: -0.0003

Data Index: 437024 (C4 (Web Text))

Max Activating Token Index: 390

Text #6

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0304. Min Act: -0.0003

Data Index: 1521634 (Python Code)

Max Activating Token Index: 1023

Text #7

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0330. Min Act: -0.0003

Data Index: 422549 (C4 (Web Text))

Max Activating Token Index: 1023

Text #8

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0324. Min Act: -0.0003

Data Index: 1437614 (Python Code)

Max Activating Token Index: 1023

Text #9

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0304. Min Act: -0.0003

Data Index: 371046 (C4 (Web Text))

Max Activating Token Index: 1023

Text #10

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0353. Min Act: -0.0003

Data Index: 537726 (C4 (Web Text))

Max Activating Token Index: 242

Text #11

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0262. Min Act: -0.0003

Data Index: 1566389 (Python Code)

Max Activating Token Index: 1023

Text #12

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0316. Min Act: -0.0003

Data Index: 1410001 (Python Code)

Max Activating Token Index: 1023

Text #13

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0295. Min Act: -0.0003

Data Index: 214850 (C4 (Web Text))

Max Activating Token Index: 1023

Text #14

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0263. Min Act: -0.0003

Data Index: 458340 (C4 (Web Text))

Max Activating Token Index: 277

Text #15

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0298. Min Act: -0.0003

Data Index: 1635305 (Python Code)

Max Activating Token Index: 4

Text #16

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0288. Min Act: -0.0003

Data Index: 1375101 (Python Code)

Max Activating Token Index: 1023

Text #17

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0291. Min Act: -0.0003

Data Index: 1195869 (C4 (Web Text))

Max Activating Token Index: 1023

Text #18

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0286. Min Act: -0.0003

Data Index: 213521 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 0.0357. Min Range: -0.0357

Max Act: 0.0255. Min Act: -0.0003

Data Index: 218107 (C4 (Web Text))

Max Activating Token Index: 1023