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

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

Neuron 671 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.8075. Min Act: -0.0001

Data Index: 1117286 (C4 (Web Text))

Max Activating Token Index: 570

Text #1

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.7046. Min Act: -0.0001

Data Index: 1596000 (Python Code)

Max Activating Token Index: 372

Text #2

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.9244. Min Act: -0.0001

Data Index: 1442825 (Python Code)

Max Activating Token Index: 732

Text #3

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.3376. Min Act: -0.0001

Data Index: 40769 (C4 (Web Text))

Max Activating Token Index: 667

Text #4

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.1796. Min Act: -0.0001

Data Index: 644132 (C4 (Web Text))

Max Activating Token Index: 4

Text #5

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.8177. Min Act: -0.0001

Data Index: 1178934 (C4 (Web Text))

Max Activating Token Index: 869

Text #6

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.9014. Min Act: -0.0001

Data Index: 935292 (C4 (Web Text))

Max Activating Token Index: 408

Text #7

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.4514. Min Act: -0.0001

Data Index: 1331366 (C4 (Web Text))

Max Activating Token Index: 410

Text #8

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.8904. Min Act: -0.0001

Data Index: 1392640 (Python Code)

Max Activating Token Index: 629

Text #9

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.1781. Min Act: -0.0001

Data Index: 1379499 (Python Code)

Max Activating Token Index: 166

Text #10

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.7175. Min Act: -0.0001

Data Index: 495883 (C4 (Web Text))

Max Activating Token Index: 274

Text #11

Max Range: 3.9244. Min Range: -3.9244

Max Act: 3.1603. Min Act: -0.0001

Data Index: 665362 (C4 (Web Text))

Max Activating Token Index: 916

Text #12

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.5333. Min Act: -0.0001

Data Index: 476793 (C4 (Web Text))

Max Activating Token Index: 243

Text #13

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.5401. Min Act: -0.0001

Data Index: 291645 (C4 (Web Text))

Max Activating Token Index: 422

Text #14

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.6731. Min Act: -0.0001

Data Index: 337409 (C4 (Web Text))

Max Activating Token Index: 332

Text #15

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.4839. Min Act: -0.0001

Data Index: 1466542 (Python Code)

Max Activating Token Index: 878

Text #16

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.5443. Min Act: -0.0001

Data Index: 681697 (C4 (Web Text))

Max Activating Token Index: 430

Text #17

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.6655. Min Act: -0.0001

Data Index: 330699 (C4 (Web Text))

Max Activating Token Index: 318

Text #18

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.6295. Min Act: -0.0001

Data Index: 3616 (C4 (Web Text))

Max Activating Token Index: 318

Text #19

Max Range: 3.9244. Min Range: -3.9244

Max Act: 2.7527. Min Act: -0.0001

Data Index: 982579 (C4 (Web Text))

Max Activating Token Index: 513