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

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

Neuron 2678 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.3879. Min Act: -0.0000

Data Index: 738038 (C4 (Web Text))

Max Activating Token Index: 362

Text #1

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.4831. Min Act: -0.0000

Data Index: 1426691 (Python Code)

Max Activating Token Index: 124

Text #2

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.3795. Min Act: -0.0000

Data Index: 443371 (C4 (Web Text))

Max Activating Token Index: 984

Text #3

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.2340. Min Act: -0.0000

Data Index: 1340618 (C4 (Web Text))

Max Activating Token Index: 531

Text #4

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.2302. Min Act: -0.0000

Data Index: 537447 (C4 (Web Text))

Max Activating Token Index: 74

Text #5

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0924. Min Act: -0.0000

Data Index: 712969 (C4 (Web Text))

Max Activating Token Index: 133

Text #6

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.1300. Min Act: -0.0000

Data Index: 869390 (C4 (Web Text))

Max Activating Token Index: 967

Text #7

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0669. Min Act: -0.0000

Data Index: 443513 (C4 (Web Text))

Max Activating Token Index: 613

Text #8

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.1845. Min Act: -0.0000

Data Index: 316975 (C4 (Web Text))

Max Activating Token Index: 215

Text #9

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0846. Min Act: -0.0001

Data Index: 1109485 (C4 (Web Text))

Max Activating Token Index: 784

Text #10

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0697. Min Act: -0.0000

Data Index: 548970 (C4 (Web Text))

Max Activating Token Index: 905

Text #11

Max Range: 1.4831. Min Range: -1.4831

Max Act: 0.9964. Min Act: -0.0000

Data Index: 88474 (C4 (Web Text))

Max Activating Token Index: 247

Text #12

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.1022. Min Act: -0.0000

Data Index: 1022951 (C4 (Web Text))

Max Activating Token Index: 904

Text #13

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0252. Min Act: -0.0000

Data Index: 1352068 (C4 (Web Text))

Max Activating Token Index: 419

Text #14

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0211. Min Act: -0.0000

Data Index: 169680 (C4 (Web Text))

Max Activating Token Index: 89

Text #15

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0129. Min Act: -0.0000

Data Index: 704717 (C4 (Web Text))

Max Activating Token Index: 63

Text #16

Max Range: 1.4831. Min Range: -1.4831

Max Act: 1.0802. Min Act: -0.0000

Data Index: 1601857 (Python Code)

Max Activating Token Index: 228

Text #17

Max Range: 1.4831. Min Range: -1.4831

Max Act: 0.9943. Min Act: -0.0000

Data Index: 1152745 (C4 (Web Text))

Max Activating Token Index: 28

Text #18

Max Range: 1.4831. Min Range: -1.4831

Max Act: 0.9317. Min Act: -0.0000

Data Index: 227982 (C4 (Web Text))

Max Activating Token Index: 640

Text #19

Max Range: 1.4831. Min Range: -1.4831

Max Act: 0.8720. Min Act: -0.0001

Data Index: 1249602 (C4 (Web Text))

Max Activating Token Index: 124