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

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

Neuron 3124 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.3052. Min Act: -0.0002

Data Index: 1446832 (Python Code)

Max Activating Token Index: 163

Text #1

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2857. Min Act: -0.0002

Data Index: 189047 (C4 (Web Text))

Max Activating Token Index: 437

Text #2

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.3609. Min Act: -0.0002

Data Index: 557391 (C4 (Web Text))

Max Activating Token Index: 699

Text #3

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2866. Min Act: -0.0002

Data Index: 1223857 (C4 (Web Text))

Max Activating Token Index: 164

Text #4

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2664. Min Act: -0.0002

Data Index: 283501 (C4 (Web Text))

Max Activating Token Index: 28

Text #5

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2841. Min Act: -0.0002

Data Index: 106454 (C4 (Web Text))

Max Activating Token Index: 22

Text #6

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2807. Min Act: -0.0002

Data Index: 1622483 (Python Code)

Max Activating Token Index: 734

Text #7

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2411. Min Act: -0.0002

Data Index: 1563666 (Python Code)

Max Activating Token Index: 1022

Text #8

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2429. Min Act: -0.0002

Data Index: 59175 (C4 (Web Text))

Max Activating Token Index: 517

Text #9

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2635. Min Act: -0.0002

Data Index: 940360 (C4 (Web Text))

Max Activating Token Index: 978

Text #10

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2302. Min Act: -0.0002

Data Index: 233136 (C4 (Web Text))

Max Activating Token Index: 888

Text #11

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2549. Min Act: -0.0002

Data Index: 758425 (C4 (Web Text))

Max Activating Token Index: 924

Text #12

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2366. Min Act: -0.0002

Data Index: 854034 (C4 (Web Text))

Max Activating Token Index: 28

Text #13

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2458. Min Act: -0.0002

Data Index: 1272503 (C4 (Web Text))

Max Activating Token Index: 12

Text #14

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2518. Min Act: -0.0002

Data Index: 169012 (C4 (Web Text))

Max Activating Token Index: 23

Text #15

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2341. Min Act: -0.0002

Data Index: 1501365 (Python Code)

Max Activating Token Index: 629

Text #16

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.1941. Min Act: -0.0002

Data Index: 1408921 (Python Code)

Max Activating Token Index: 726

Text #17

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2196. Min Act: -0.0002

Data Index: 1203435 (C4 (Web Text))

Max Activating Token Index: 941

Text #18

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2309. Min Act: -0.0002

Data Index: 367342 (C4 (Web Text))

Max Activating Token Index: 743

Text #19

Max Range: 0.3609. Min Range: -0.3609

Max Act: 0.2045. Min Act: -0.0002

Data Index: 86543 (C4 (Web Text))

Max Activating Token Index: 319