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

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

Neuron 1825 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0777. Min Act: -0.0001

Data Index: 244919 (C4 (Web Text))

Max Activating Token Index: 289

Text #1

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0498. Min Act: -0.0001

Data Index: 528801 (C4 (Web Text))

Max Activating Token Index: 668

Text #2

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0465. Min Act: -0.0001

Data Index: 1223894 (C4 (Web Text))

Max Activating Token Index: 749

Text #3

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0424. Min Act: -0.0001

Data Index: 463337 (C4 (Web Text))

Max Activating Token Index: 1016

Text #4

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0421. Min Act: -0.0001

Data Index: 652344 (C4 (Web Text))

Max Activating Token Index: 643

Text #5

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0400. Min Act: -0.0001

Data Index: 290273 (C4 (Web Text))

Max Activating Token Index: 327

Text #6

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0398. Min Act: -0.0001

Data Index: 104979 (C4 (Web Text))

Max Activating Token Index: 395

Text #7

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0421. Min Act: -0.0001

Data Index: 657637 (C4 (Web Text))

Max Activating Token Index: 459

Text #8

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0389. Min Act: -0.0001

Data Index: 79176 (C4 (Web Text))

Max Activating Token Index: 960

Text #9

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0357. Min Act: -0.0001

Data Index: 1127505 (C4 (Web Text))

Max Activating Token Index: 680

Text #10

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0377. Min Act: -0.0001

Data Index: 697945 (C4 (Web Text))

Max Activating Token Index: 914

Text #11

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0400. Min Act: -0.0001

Data Index: 625848 (C4 (Web Text))

Max Activating Token Index: 601

Text #12

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0346. Min Act: -0.0001

Data Index: 1408960 (Python Code)

Max Activating Token Index: 301

Text #13

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0388. Min Act: -0.0001

Data Index: 868901 (C4 (Web Text))

Max Activating Token Index: 894

Text #14

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0368. Min Act: -0.0001

Data Index: 1300535 (C4 (Web Text))

Max Activating Token Index: 554

Text #15

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0375. Min Act: -0.0001

Data Index: 201557 (C4 (Web Text))

Max Activating Token Index: 558

Text #16

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0378. Min Act: -0.0001

Data Index: 190546 (C4 (Web Text))

Max Activating Token Index: 880

Text #17

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0370. Min Act: -0.0001

Data Index: 381745 (C4 (Web Text))

Max Activating Token Index: 824

Text #18

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0367. Min Act: -0.0001

Data Index: 922578 (C4 (Web Text))

Max Activating Token Index: 830

Text #19

Max Range: 0.0777. Min Range: -0.0777

Max Act: 0.0347. Min Act: -0.0001

Data Index: 357140 (C4 (Web Text))

Max Activating Token Index: 828