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

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

Neuron 655 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2849. Min Act: -0.0000

Data Index: 1176205 (C4 (Web Text))

Max Activating Token Index: 232

Text #1

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2628. Min Act: -0.0000

Data Index: 1059280 (C4 (Web Text))

Max Activating Token Index: 273

Text #2

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2580. Min Act: -0.0001

Data Index: 1078312 (C4 (Web Text))

Max Activating Token Index: 83

Text #3

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2451. Min Act: -0.0000

Data Index: 634263 (C4 (Web Text))

Max Activating Token Index: 467

Text #4

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2544. Min Act: -0.0000

Data Index: 544611 (C4 (Web Text))

Max Activating Token Index: 510

Text #5

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2237. Min Act: -0.0001

Data Index: 831908 (C4 (Web Text))

Max Activating Token Index: 268

Text #6

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1957. Min Act: -0.0000

Data Index: 771314 (C4 (Web Text))

Max Activating Token Index: 811

Text #7

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1814. Min Act: -0.0001

Data Index: 805502 (C4 (Web Text))

Max Activating Token Index: 517

Text #8

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2030. Min Act: -0.0000

Data Index: 1125427 (C4 (Web Text))

Max Activating Token Index: 163

Text #9

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.2072. Min Act: -0.0001

Data Index: 864043 (C4 (Web Text))

Max Activating Token Index: 650

Text #10

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1818. Min Act: -0.0000

Data Index: 151221 (C4 (Web Text))

Max Activating Token Index: 215

Text #11

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1757. Min Act: -0.0000

Data Index: 634856 (C4 (Web Text))

Max Activating Token Index: 474

Text #12

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1593. Min Act: -0.0001

Data Index: 1495749 (Python Code)

Max Activating Token Index: 740

Text #13

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1636. Min Act: -0.0001

Data Index: 1505002 (Python Code)

Max Activating Token Index: 487

Text #14

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1662. Min Act: -0.0000

Data Index: 498456 (C4 (Web Text))

Max Activating Token Index: 178

Text #15

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1533. Min Act: -0.0001

Data Index: 1433626 (Python Code)

Max Activating Token Index: 281

Text #16

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1504. Min Act: -0.0001

Data Index: 1631395 (Python Code)

Max Activating Token Index: 861

Text #17

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1531. Min Act: -0.0001

Data Index: 1402110 (Python Code)

Max Activating Token Index: 571

Text #18

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1513. Min Act: -0.0001

Data Index: 1537249 (Python Code)

Max Activating Token Index: 648

Text #19

Max Range: 0.2849. Min Range: -0.2849

Max Act: 0.1372. Min Act: -0.0001

Data Index: 1475168 (Python Code)

Max Activating Token Index: 929