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

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

Neuron 3287 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.1800. Min Range: -4.1800

Max Act: 4.1800. Min Act: -0.0000

Data Index: 1377925 (Python Code)

Max Activating Token Index: 590

Text #1

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.7252. Min Act: -0.0000

Data Index: 389773 (C4 (Web Text))

Max Activating Token Index: 774

Text #2

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.8739. Min Act: -0.0000

Data Index: 924093 (C4 (Web Text))

Max Activating Token Index: 484

Text #3

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.6472. Min Act: -0.0000

Data Index: 1273437 (C4 (Web Text))

Max Activating Token Index: 373

Text #4

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.6016. Min Act: -0.0000

Data Index: 367030 (C4 (Web Text))

Max Activating Token Index: 916

Text #5

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.7615. Min Act: -0.0000

Data Index: 1348063 (C4 (Web Text))

Max Activating Token Index: 9

Text #6

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.7713. Min Act: -0.0000

Data Index: 1057572 (C4 (Web Text))

Max Activating Token Index: 323

Text #7

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.7088. Min Act: -0.0000

Data Index: 788123 (C4 (Web Text))

Max Activating Token Index: 16

Text #8

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.5120. Min Act: -0.0000

Data Index: 944557 (C4 (Web Text))

Max Activating Token Index: 708

Text #9

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.3984. Min Act: -0.0000

Data Index: 906512 (C4 (Web Text))

Max Activating Token Index: 833

Text #10

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.4068. Min Act: -0.0000

Data Index: 1311264 (C4 (Web Text))

Max Activating Token Index: 891

Text #11

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.2874. Min Act: -0.0000

Data Index: 1245712 (C4 (Web Text))

Max Activating Token Index: 576

Text #12

Max Range: 4.1800. Min Range: -4.1800

Max Act: 2.9470. Min Act: -0.0000

Data Index: 520361 (C4 (Web Text))

Max Activating Token Index: 855

Text #13

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.4908. Min Act: -0.0000

Data Index: 598010 (C4 (Web Text))

Max Activating Token Index: 330

Text #14

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.0785. Min Act: -0.0000

Data Index: 17696 (C4 (Web Text))

Max Activating Token Index: 559

Text #15

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.1797. Min Act: -0.0000

Data Index: 903820 (C4 (Web Text))

Max Activating Token Index: 184

Text #16

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.0294. Min Act: -0.0000

Data Index: 643604 (C4 (Web Text))

Max Activating Token Index: 393

Text #17

Max Range: 4.1800. Min Range: -4.1800

Max Act: 2.9974. Min Act: -0.0000

Data Index: 850905 (C4 (Web Text))

Max Activating Token Index: 569

Text #18

Max Range: 4.1800. Min Range: -4.1800

Max Act: 2.9002. Min Act: -0.0000

Data Index: 109391 (C4 (Web Text))

Max Activating Token Index: 623

Text #19

Max Range: 4.1800. Min Range: -4.1800

Max Act: 3.0868. Min Act: -0.0000

Data Index: 1394796 (Python Code)

Max Activating Token Index: 721