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

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

Neuron 809 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6268. Min Act: -0.0001

Data Index: 601805 (C4 (Web Text))

Max Activating Token Index: 296

Text #1

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.8233. Min Act: -0.0001

Data Index: 1516210 (Python Code)

Max Activating Token Index: 739

Text #2

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7881. Min Act: -0.0001

Data Index: 1572931 (Python Code)

Max Activating Token Index: 835

Text #3

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.8418. Min Act: -0.0001

Data Index: 604997 (C4 (Web Text))

Max Activating Token Index: 570

Text #4

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6556. Min Act: -0.0001

Data Index: 1291335 (C4 (Web Text))

Max Activating Token Index: 739

Text #5

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7164. Min Act: -0.0001

Data Index: 521933 (C4 (Web Text))

Max Activating Token Index: 663

Text #6

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.5912. Min Act: -0.0001

Data Index: 579558 (C4 (Web Text))

Max Activating Token Index: 412

Text #7

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6004. Min Act: -0.0001

Data Index: 1592500 (Python Code)

Max Activating Token Index: 137

Text #8

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6692. Min Act: -0.0001

Data Index: 647536 (C4 (Web Text))

Max Activating Token Index: 467

Text #9

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7580. Min Act: -0.0001

Data Index: 1329065 (C4 (Web Text))

Max Activating Token Index: 788

Text #10

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7608. Min Act: -0.0001

Data Index: 1272505 (C4 (Web Text))

Max Activating Token Index: 987

Text #11

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7843. Min Act: -0.0001

Data Index: 233318 (C4 (Web Text))

Max Activating Token Index: 728

Text #12

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6671. Min Act: -0.0000

Data Index: 1436528 (Python Code)

Max Activating Token Index: 101

Text #13

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6876. Min Act: -0.0001

Data Index: 522006 (C4 (Web Text))

Max Activating Token Index: 98

Text #14

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.7627. Min Act: -0.0000

Data Index: 1605360 (Python Code)

Max Activating Token Index: 125

Text #15

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6608. Min Act: -0.0001

Data Index: 173327 (C4 (Web Text))

Max Activating Token Index: 752

Text #16

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6260. Min Act: -0.0001

Data Index: 1143146 (C4 (Web Text))

Max Activating Token Index: 813

Text #17

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.6355. Min Act: -0.0001

Data Index: 1308875 (C4 (Web Text))

Max Activating Token Index: 594

Text #18

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.5377. Min Act: -0.0001

Data Index: 202038 (C4 (Web Text))

Max Activating Token Index: 66

Text #19

Max Range: 1.8418. Min Range: -1.8418

Max Act: 1.5851. Min Act: -0.0001

Data Index: 1485534 (Python Code)

Max Activating Token Index: 501