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

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

Neuron 4929 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.2617. Min Act: -0.0000

Data Index: 55076 (C4 (Web Text))

Max Activating Token Index: 266

Text #1

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.2306. Min Act: -0.0000

Data Index: 369160 (C4 (Web Text))

Max Activating Token Index: 143

Text #2

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.1592. Min Act: -0.0000

Data Index: 862751 (C4 (Web Text))

Max Activating Token Index: 609

Text #3

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0985. Min Act: -0.0000

Data Index: 1138815 (C4 (Web Text))

Max Activating Token Index: 927

Text #4

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.1528. Min Act: -0.0000

Data Index: 13793 (C4 (Web Text))

Max Activating Token Index: 461

Text #5

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0381. Min Act: -0.0000

Data Index: 1317988 (C4 (Web Text))

Max Activating Token Index: 739

Text #6

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.2092. Min Act: -0.0000

Data Index: 613908 (C4 (Web Text))

Max Activating Token Index: 74

Text #7

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0609. Min Act: -0.0000

Data Index: 1638416 (Python Code)

Max Activating Token Index: 467

Text #8

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0567. Min Act: -0.0000

Data Index: 3447 (C4 (Web Text))

Max Activating Token Index: 858

Text #9

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0828. Min Act: -0.0000

Data Index: 1196237 (C4 (Web Text))

Max Activating Token Index: 249

Text #10

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.1188. Min Act: -0.0000

Data Index: 296067 (C4 (Web Text))

Max Activating Token Index: 539

Text #11

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0489. Min Act: -0.0000

Data Index: 1242495 (C4 (Web Text))

Max Activating Token Index: 50

Text #12

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0156. Min Act: -0.0000

Data Index: 367355 (C4 (Web Text))

Max Activating Token Index: 913

Text #13

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0592. Min Act: -0.0000

Data Index: 229830 (C4 (Web Text))

Max Activating Token Index: 40

Text #14

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0653. Min Act: -0.0000

Data Index: 55144 (C4 (Web Text))

Max Activating Token Index: 16

Text #15

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0300. Min Act: -0.0000

Data Index: 1587852 (Python Code)

Max Activating Token Index: 469

Text #16

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0532. Min Act: -0.0000

Data Index: 642587 (C4 (Web Text))

Max Activating Token Index: 1014

Text #17

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0182. Min Act: -0.0000

Data Index: 1363140 (Python Code)

Max Activating Token Index: 576

Text #18

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0482. Min Act: -0.0000

Data Index: 1099690 (C4 (Web Text))

Max Activating Token Index: 950

Text #19

Max Range: 1.2617. Min Range: -1.2617

Max Act: 1.0084. Min Act: -0.0000

Data Index: 763566 (C4 (Web Text))

Max Activating Token Index: 736