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

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

Neuron 12 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1494. Min Act: -0.0001

Data Index: 187905 (C4 (Web Text))

Max Activating Token Index: 571

Text #1

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1444. Min Act: -0.0001

Data Index: 160409 (C4 (Web Text))

Max Activating Token Index: 654

Text #2

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1485. Min Act: -0.0001

Data Index: 1656483 (Python Code)

Max Activating Token Index: 134

Text #3

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1330. Min Act: -0.0001

Data Index: 176678 (C4 (Web Text))

Max Activating Token Index: 285

Text #4

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1416. Min Act: -0.0001

Data Index: 1515968 (Python Code)

Max Activating Token Index: 27

Text #5

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1472. Min Act: -0.0001

Data Index: 305544 (C4 (Web Text))

Max Activating Token Index: 602

Text #6

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1333. Min Act: -0.0001

Data Index: 1270530 (C4 (Web Text))

Max Activating Token Index: 1018

Text #7

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1366. Min Act: -0.0001

Data Index: 741607 (C4 (Web Text))

Max Activating Token Index: 57

Text #8

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1284. Min Act: -0.0001

Data Index: 398701 (C4 (Web Text))

Max Activating Token Index: 253

Text #9

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1377. Min Act: -0.0001

Data Index: 1555265 (Python Code)

Max Activating Token Index: 257

Text #10

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1556. Min Act: -0.0001

Data Index: 517728 (C4 (Web Text))

Max Activating Token Index: 250

Text #11

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1297. Min Act: -0.0001

Data Index: 214848 (C4 (Web Text))

Max Activating Token Index: 95

Text #12

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1303. Min Act: -0.0001

Data Index: 1469432 (Python Code)

Max Activating Token Index: 248

Text #13

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1352. Min Act: -0.0001

Data Index: 1195617 (C4 (Web Text))

Max Activating Token Index: 368

Text #14

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1432. Min Act: -0.0001

Data Index: 1435695 (Python Code)

Max Activating Token Index: 849

Text #15

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1475. Min Act: -0.0001

Data Index: 417384 (C4 (Web Text))

Max Activating Token Index: 98

Text #16

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1360. Min Act: -0.0001

Data Index: 170963 (C4 (Web Text))

Max Activating Token Index: 358

Text #17

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1312. Min Act: -0.0001

Data Index: 1387438 (Python Code)

Max Activating Token Index: 357

Text #18

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1309. Min Act: -0.0001

Data Index: 1430348 (Python Code)

Max Activating Token Index: 6

Text #19

Max Range: 0.1556. Min Range: -0.1556

Max Act: 0.1195. Min Act: -0.0001

Data Index: 811959 (C4 (Web Text))

Max Activating Token Index: 110