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

Dataset: The Pile

Neuron 64 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.4037. Min Act: -0.0000

Data Index: 1841612 (The Pile)

Max Activating Token Index: 495

Text #1

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.0085. Min Act: -0.0000

Data Index: 1632084 (The Pile)

Max Activating Token Index: 885

Text #2

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.6577. Min Act: -0.0000

Data Index: 40448 (The Pile)

Max Activating Token Index: 589

Text #3

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.1916. Min Act: -0.0000

Data Index: 212005 (The Pile)

Max Activating Token Index: 518

Text #4

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.2601. Min Act: -0.0000

Data Index: 184533 (The Pile)

Max Activating Token Index: 982

Text #5

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.3812. Min Act: -0.0000

Data Index: 787687 (The Pile)

Max Activating Token Index: 552

Text #6

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.3724. Min Act: -0.0000

Data Index: 396611 (The Pile)

Max Activating Token Index: 547

Text #7

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.2217. Min Act: -0.0000

Data Index: 1205216 (The Pile)

Max Activating Token Index: 1006

Text #8

Max Range: 5.6577. Min Range: -5.6577

Max Act: 5.1098. Min Act: -0.0000

Data Index: 1433846 (The Pile)

Max Activating Token Index: 880

Text #9

Max Range: 5.6577. Min Range: -5.6577

Max Act: 4.6345. Min Act: -0.0000

Data Index: 2787 (The Pile)

Max Activating Token Index: 502

Text #10

Max Range: 5.6577. Min Range: -5.6577

Max Act: 4.2811. Min Act: -0.0000

Data Index: 683601 (The Pile)

Max Activating Token Index: 118

Text #11

Max Range: 5.6577. Min Range: -5.6577

Max Act: 4.4234. Min Act: -0.0000

Data Index: 496302 (The Pile)

Max Activating Token Index: 137

Text #12

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.7286. Min Act: -0.0000

Data Index: 54512 (The Pile)

Max Activating Token Index: 258

Text #13

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.9517. Min Act: -0.0000

Data Index: 477735 (The Pile)

Max Activating Token Index: 964

Text #14

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.4537. Min Act: -0.0000

Data Index: 172526 (The Pile)

Max Activating Token Index: 63

Text #15

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.8636. Min Act: -0.0000

Data Index: 1538048 (The Pile)

Max Activating Token Index: 388

Text #16

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.1999. Min Act: -0.0000

Data Index: 1485083 (The Pile)

Max Activating Token Index: 231

Text #17

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.3564. Min Act: -0.0000

Data Index: 731190 (The Pile)

Max Activating Token Index: 959

Text #18

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.1663. Min Act: -0.0000

Data Index: 1461404 (The Pile)

Max Activating Token Index: 780

Text #19

Max Range: 5.6577. Min Range: -5.6577

Max Act: 3.3313. Min Act: -0.0000

Data Index: 1120113 (The Pile)

Max Activating Token Index: 220