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

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

Neuron 4807 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.3519. Min Act: -0.0000

Data Index: 77283 (C4 (Web Text))

Max Activating Token Index: 695

Text #1

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.2886. Min Act: -0.0000

Data Index: 748707 (C4 (Web Text))

Max Activating Token Index: 196

Text #2

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.2592. Min Act: -0.0000

Data Index: 331695 (C4 (Web Text))

Max Activating Token Index: 856

Text #3

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.2490. Min Act: -0.0000

Data Index: 749241 (C4 (Web Text))

Max Activating Token Index: 417

Text #4

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.2300. Min Act: -0.0000

Data Index: 746476 (C4 (Web Text))

Max Activating Token Index: 828

Text #5

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1736. Min Act: -0.0000

Data Index: 646646 (C4 (Web Text))

Max Activating Token Index: 176

Text #6

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1785. Min Act: -0.0000

Data Index: 1578275 (Python Code)

Max Activating Token Index: 758

Text #7

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1762. Min Act: -0.0000

Data Index: 251106 (C4 (Web Text))

Max Activating Token Index: 230

Text #8

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1860. Min Act: -0.0000

Data Index: 291281 (C4 (Web Text))

Max Activating Token Index: 746

Text #9

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1879. Min Act: -0.0000

Data Index: 745451 (C4 (Web Text))

Max Activating Token Index: 708

Text #10

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1690. Min Act: -0.0000

Data Index: 1572663 (Python Code)

Max Activating Token Index: 618

Text #11

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1739. Min Act: -0.0000

Data Index: 1481632 (Python Code)

Max Activating Token Index: 289

Text #12

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1585. Min Act: -0.0000

Data Index: 666517 (C4 (Web Text))

Max Activating Token Index: 753

Text #13

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1575. Min Act: -0.0000

Data Index: 1346467 (C4 (Web Text))

Max Activating Token Index: 979

Text #14

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1558. Min Act: -0.0000

Data Index: 1595577 (Python Code)

Max Activating Token Index: 518

Text #15

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1383. Min Act: -0.0000

Data Index: 668068 (C4 (Web Text))

Max Activating Token Index: 839

Text #16

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1650. Min Act: -0.0000

Data Index: 1595552 (Python Code)

Max Activating Token Index: 498

Text #17

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1456. Min Act: -0.0000

Data Index: 1162598 (C4 (Web Text))

Max Activating Token Index: 91

Text #18

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1600. Min Act: -0.0000

Data Index: 674473 (C4 (Web Text))

Max Activating Token Index: 120

Text #19

Max Range: 0.3519. Min Range: -0.3519

Max Act: 0.1465. Min Act: -0.0000

Data Index: 406738 (C4 (Web Text))

Max Activating Token Index: 246