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

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

Neuron 3789 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1873. Min Act: -0.0001

Data Index: 815392 (C4 (Web Text))

Max Activating Token Index: 49

Text #1

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1729. Min Act: -0.0002

Data Index: 76894 (C4 (Web Text))

Max Activating Token Index: 294

Text #2

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1746. Min Act: -0.0002

Data Index: 552877 (C4 (Web Text))

Max Activating Token Index: 756

Text #3

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1557. Min Act: -0.0001

Data Index: 1608620 (Python Code)

Max Activating Token Index: 461

Text #4

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1602. Min Act: -0.0002

Data Index: 172306 (C4 (Web Text))

Max Activating Token Index: 504

Text #5

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1545. Min Act: -0.0002

Data Index: 113102 (C4 (Web Text))

Max Activating Token Index: 2

Text #6

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1469. Min Act: -0.0002

Data Index: 490176 (C4 (Web Text))

Max Activating Token Index: 962

Text #7

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1467. Min Act: -0.0001

Data Index: 364499 (C4 (Web Text))

Max Activating Token Index: 27

Text #8

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1403. Min Act: -0.0001

Data Index: 558863 (C4 (Web Text))

Max Activating Token Index: 29

Text #9

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1483. Min Act: -0.0002

Data Index: 1196304 (C4 (Web Text))

Max Activating Token Index: 484

Text #10

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1530. Min Act: -0.0002

Data Index: 265250 (C4 (Web Text))

Max Activating Token Index: 657

Text #11

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1434. Min Act: -0.0001

Data Index: 1528438 (Python Code)

Max Activating Token Index: 691

Text #12

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1462. Min Act: -0.0001

Data Index: 1205568 (C4 (Web Text))

Max Activating Token Index: 520

Text #13

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1399. Min Act: -0.0001

Data Index: 203900 (C4 (Web Text))

Max Activating Token Index: 644

Text #14

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1338. Min Act: -0.0002

Data Index: 1343587 (C4 (Web Text))

Max Activating Token Index: 355

Text #15

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1288. Min Act: -0.0002

Data Index: 818927 (C4 (Web Text))

Max Activating Token Index: 606

Text #16

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1352. Min Act: -0.0001

Data Index: 161601 (C4 (Web Text))

Max Activating Token Index: 335

Text #17

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1331. Min Act: -0.0001

Data Index: 761287 (C4 (Web Text))

Max Activating Token Index: 594

Text #18

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1272. Min Act: -0.0001

Data Index: 1361896 (Python Code)

Max Activating Token Index: 760

Text #19

Max Range: 0.1873. Min Range: -0.1873

Max Act: 0.1313. Min Act: -0.0001

Data Index: 1200080 (C4 (Web Text))

Max Activating Token Index: 51