Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 94 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0838. Min Act: -0.0002

Data Index: 780295 (C4 (Web Text))

Max Activating Token Index: 372

Text #1

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0801. Min Act: -0.0002

Data Index: 1116459 (C4 (Web Text))

Max Activating Token Index: 620

Text #2

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0824. Min Act: -0.0001

Data Index: 657161 (C4 (Web Text))

Max Activating Token Index: 885

Text #3

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0804. Min Act: -0.0003

Data Index: 928530 (C4 (Web Text))

Max Activating Token Index: 205

Text #4

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0797. Min Act: -0.0002

Data Index: 664150 (C4 (Web Text))

Max Activating Token Index: 328

Text #5

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0754. Min Act: -0.0002

Data Index: 1422246 (Python Code)

Max Activating Token Index: 827

Text #6

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0766. Min Act: -0.0002

Data Index: 848277 (C4 (Web Text))

Max Activating Token Index: 710

Text #7

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0795. Min Act: -0.0001

Data Index: 1321102 (C4 (Web Text))

Max Activating Token Index: 750

Text #8

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0752. Min Act: -0.0002

Data Index: 1215552 (C4 (Web Text))

Max Activating Token Index: 525

Text #9

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0724. Min Act: -0.0002

Data Index: 674299 (C4 (Web Text))

Max Activating Token Index: 898

Text #10

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0693. Min Act: -0.0002

Data Index: 1221769 (C4 (Web Text))

Max Activating Token Index: 367

Text #11

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0745. Min Act: -0.0001

Data Index: 1331091 (C4 (Web Text))

Max Activating Token Index: 358

Text #12

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0692. Min Act: -0.0001

Data Index: 935129 (C4 (Web Text))

Max Activating Token Index: 219

Text #13

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0717. Min Act: -0.0002

Data Index: 575756 (C4 (Web Text))

Max Activating Token Index: 619

Text #14

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0720. Min Act: -0.0003

Data Index: 1550975 (Python Code)

Max Activating Token Index: 212

Text #15

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0716. Min Act: -0.0002

Data Index: 936756 (C4 (Web Text))

Max Activating Token Index: 226

Text #16

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0683. Min Act: -0.0003

Data Index: 1510670 (Python Code)

Max Activating Token Index: 862

Text #17

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0691. Min Act: -0.0002

Data Index: 130885 (C4 (Web Text))

Max Activating Token Index: 905

Text #18

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0692. Min Act: -0.0002

Data Index: 913635 (C4 (Web Text))

Max Activating Token Index: 426

Text #19

Max Range: 0.0838. Min Range: -0.0838

Max Act: 0.0662. Min Act: -0.0002

Data Index: 535353 (C4 (Web Text))

Max Activating Token Index: 962