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

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

Neuron 2952 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: 0.4752. Min Range: -0.4752

Max Act: 0.4752. Min Act: -0.0000

Data Index: 1655971 (Python Code)

Max Activating Token Index: 643

Text #1

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.4539. Min Act: -0.0000

Data Index: 810427 (C4 (Web Text))

Max Activating Token Index: 511

Text #2

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.4356. Min Act: -0.0000

Data Index: 1045364 (C4 (Web Text))

Max Activating Token Index: 160

Text #3

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.4323. Min Act: -0.0000

Data Index: 1236448 (C4 (Web Text))

Max Activating Token Index: 49

Text #4

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3878. Min Act: -0.0000

Data Index: 1553994 (Python Code)

Max Activating Token Index: 997

Text #5

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3878. Min Act: -0.0000

Data Index: 1472821 (Python Code)

Max Activating Token Index: 965

Text #6

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3465. Min Act: -0.0000

Data Index: 658364 (C4 (Web Text))

Max Activating Token Index: 860

Text #7

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3534. Min Act: -0.0000

Data Index: 576017 (C4 (Web Text))

Max Activating Token Index: 372

Text #8

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3465. Min Act: -0.0000

Data Index: 940259 (C4 (Web Text))

Max Activating Token Index: 588

Text #9

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3500. Min Act: -0.0000

Data Index: 1601146 (Python Code)

Max Activating Token Index: 462

Text #10

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3277. Min Act: -0.0000

Data Index: 1451158 (Python Code)

Max Activating Token Index: 240

Text #11

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3420. Min Act: -0.0000

Data Index: 768258 (C4 (Web Text))

Max Activating Token Index: 296

Text #12

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3135. Min Act: -0.0000

Data Index: 1122746 (C4 (Web Text))

Max Activating Token Index: 442

Text #13

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3130. Min Act: -0.0000

Data Index: 1585855 (Python Code)

Max Activating Token Index: 287

Text #14

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3546. Min Act: -0.0000

Data Index: 275080 (C4 (Web Text))

Max Activating Token Index: 621

Text #15

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3370. Min Act: -0.0000

Data Index: 1319626 (C4 (Web Text))

Max Activating Token Index: 826

Text #16

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3041. Min Act: -0.0000

Data Index: 1048290 (C4 (Web Text))

Max Activating Token Index: 246

Text #17

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.2698. Min Act: -0.0000

Data Index: 178490 (C4 (Web Text))

Max Activating Token Index: 637

Text #18

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3056. Min Act: -0.0000

Data Index: 1476511 (Python Code)

Max Activating Token Index: 656

Text #19

Max Range: 0.4752. Min Range: -0.4752

Max Act: 0.3033. Min Act: -0.0000

Data Index: 367351 (C4 (Web Text))

Max Activating Token Index: 843