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

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

Neuron 1704 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.2274. Min Act: -0.0000

Data Index: 567895 (C4 (Web Text))

Max Activating Token Index: 480

Text #1

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.2366. Min Act: -0.0001

Data Index: 71264 (C4 (Web Text))

Max Activating Token Index: 173

Text #2

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.2137. Min Act: -0.0000

Data Index: 1130312 (C4 (Web Text))

Max Activating Token Index: 374

Text #3

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1992. Min Act: -0.0001

Data Index: 987688 (C4 (Web Text))

Max Activating Token Index: 849

Text #4

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1998. Min Act: -0.0001

Data Index: 663969 (C4 (Web Text))

Max Activating Token Index: 374

Text #5

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1943. Min Act: -0.0001

Data Index: 1197909 (C4 (Web Text))

Max Activating Token Index: 494

Text #6

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1874. Min Act: -0.0001

Data Index: 208848 (C4 (Web Text))

Max Activating Token Index: 282

Text #7

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1909. Min Act: -0.0001

Data Index: 940714 (C4 (Web Text))

Max Activating Token Index: 760

Text #8

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1790. Min Act: -0.0001

Data Index: 499672 (C4 (Web Text))

Max Activating Token Index: 700

Text #9

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1785. Min Act: -0.0001

Data Index: 1544267 (Python Code)

Max Activating Token Index: 560

Text #10

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1794. Min Act: -0.0001

Data Index: 725601 (C4 (Web Text))

Max Activating Token Index: 136

Text #11

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1648. Min Act: -0.0001

Data Index: 236376 (C4 (Web Text))

Max Activating Token Index: 1008

Text #12

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1777. Min Act: -0.0000

Data Index: 1024658 (C4 (Web Text))

Max Activating Token Index: 850

Text #13

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1658. Min Act: -0.0001

Data Index: 646019 (C4 (Web Text))

Max Activating Token Index: 446

Text #14

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1663. Min Act: -0.0001

Data Index: 255414 (C4 (Web Text))

Max Activating Token Index: 1018

Text #15

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1642. Min Act: -0.0001

Data Index: 560200 (C4 (Web Text))

Max Activating Token Index: 906

Text #16

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1697. Min Act: -0.0001

Data Index: 120631 (C4 (Web Text))

Max Activating Token Index: 658

Text #17

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1624. Min Act: -0.0000

Data Index: 786313 (C4 (Web Text))

Max Activating Token Index: 452

Text #18

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1668. Min Act: -0.0001

Data Index: 1469073 (Python Code)

Max Activating Token Index: 610

Text #19

Max Range: 0.2366. Min Range: -0.2366

Max Act: 0.1723. Min Act: -0.0001

Data Index: 587334 (C4 (Web Text))

Max Activating Token Index: 312