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

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

Neuron 2574 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.2150. Min Act: -0.0000

Data Index: 984338 (C4 (Web Text))

Max Activating Token Index: 673

Text #1

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1745. Min Act: -0.0000

Data Index: 691263 (C4 (Web Text))

Max Activating Token Index: 404

Text #2

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1690. Min Act: -0.0000

Data Index: 46093 (C4 (Web Text))

Max Activating Token Index: 714

Text #3

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1842. Min Act: -0.0000

Data Index: 318725 (C4 (Web Text))

Max Activating Token Index: 450

Text #4

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1958. Min Act: -0.0000

Data Index: 438838 (C4 (Web Text))

Max Activating Token Index: 250

Text #5

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1768. Min Act: -0.0000

Data Index: 122662 (C4 (Web Text))

Max Activating Token Index: 439

Text #6

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1684. Min Act: -0.0000

Data Index: 1595514 (Python Code)

Max Activating Token Index: 81

Text #7

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1503. Min Act: -0.0000

Data Index: 1256125 (C4 (Web Text))

Max Activating Token Index: 492

Text #8

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1546. Min Act: -0.0000

Data Index: 1524224 (Python Code)

Max Activating Token Index: 913

Text #9

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1440. Min Act: -0.0000

Data Index: 985283 (C4 (Web Text))

Max Activating Token Index: 828

Text #10

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1437. Min Act: -0.0000

Data Index: 61631 (C4 (Web Text))

Max Activating Token Index: 408

Text #11

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1699. Min Act: -0.0000

Data Index: 385589 (C4 (Web Text))

Max Activating Token Index: 777

Text #12

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1567. Min Act: -0.0000

Data Index: 738073 (C4 (Web Text))

Max Activating Token Index: 622

Text #13

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1269. Min Act: -0.0000

Data Index: 615365 (C4 (Web Text))

Max Activating Token Index: 188

Text #14

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1515. Min Act: -0.0000

Data Index: 1430510 (Python Code)

Max Activating Token Index: 685

Text #15

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1035. Min Act: -0.0000

Data Index: 221504 (C4 (Web Text))

Max Activating Token Index: 649

Text #16

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1399. Min Act: -0.0000

Data Index: 778624 (C4 (Web Text))

Max Activating Token Index: 739

Text #17

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1285. Min Act: -0.0000

Data Index: 572056 (C4 (Web Text))

Max Activating Token Index: 384

Text #18

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1492. Min Act: -0.0000

Data Index: 96689 (C4 (Web Text))

Max Activating Token Index: 706

Text #19

Max Range: 0.2150. Min Range: -0.2150

Max Act: 0.1285. Min Act: -0.0000

Data Index: 770232 (C4 (Web Text))

Max Activating Token Index: 302