Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2913 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.8384. Min Act: -0.0001

Data Index: 825929 (C4 (Web Text))

Max Activating Token Index: 970

Text #1

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.8291. Min Act: -0.0001

Data Index: 760733 (C4 (Web Text))

Max Activating Token Index: 137

Text #2

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.7526. Min Act: -0.0001

Data Index: 672997 (C4 (Web Text))

Max Activating Token Index: 27

Text #3

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.8078. Min Act: -0.0001

Data Index: 1535748 (Python Code)

Max Activating Token Index: 634

Text #4

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.6623. Min Act: -0.0001

Data Index: 734836 (C4 (Web Text))

Max Activating Token Index: 363

Text #5

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.6866. Min Act: -0.0001

Data Index: 662857 (C4 (Web Text))

Max Activating Token Index: 990

Text #6

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.6290. Min Act: -0.0001

Data Index: 1205986 (C4 (Web Text))

Max Activating Token Index: 112

Text #7

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.6194. Min Act: -0.0001

Data Index: 868049 (C4 (Web Text))

Max Activating Token Index: 802

Text #8

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.4338. Min Act: -0.0001

Data Index: 735918 (C4 (Web Text))

Max Activating Token Index: 616

Text #9

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.5016. Min Act: -0.0001

Data Index: 656322 (C4 (Web Text))

Max Activating Token Index: 746

Text #10

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.5668. Min Act: -0.0001

Data Index: 103030 (C4 (Web Text))

Max Activating Token Index: 72

Text #11

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.5671. Min Act: -0.0001

Data Index: 527348 (C4 (Web Text))

Max Activating Token Index: 372

Text #12

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.6350. Min Act: -0.0001

Data Index: 365206 (C4 (Web Text))

Max Activating Token Index: 756

Text #13

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.4910. Min Act: -0.0001

Data Index: 1623526 (Python Code)

Max Activating Token Index: 975

Text #14

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.4338. Min Act: -0.0001

Data Index: 596316 (C4 (Web Text))

Max Activating Token Index: 169

Text #15

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.4763. Min Act: -0.0001

Data Index: 1466855 (Python Code)

Max Activating Token Index: 183

Text #16

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.3870. Min Act: -0.0001

Data Index: 1579977 (Python Code)

Max Activating Token Index: 882

Text #17

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.5014. Min Act: -0.0001

Data Index: 605324 (C4 (Web Text))

Max Activating Token Index: 828

Text #18

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.3933. Min Act: -0.0001

Data Index: 617200 (C4 (Web Text))

Max Activating Token Index: 509

Text #19

Max Range: 1.8384. Min Range: -1.8384

Max Act: 1.3767. Min Act: -0.0001

Data Index: 1532134 (Python Code)

Max Activating Token Index: 505