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

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

Neuron 1070 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.3517. Min Act: -0.0000

Data Index: 1133504 (C4 (Web Text))

Max Activating Token Index: 383

Text #1

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.3061. Min Act: -0.0000

Data Index: 215542 (C4 (Web Text))

Max Activating Token Index: 574

Text #2

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.3158. Min Act: -0.0000

Data Index: 1155191 (C4 (Web Text))

Max Activating Token Index: 208

Text #3

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.1484. Min Act: -0.0000

Data Index: 258158 (C4 (Web Text))

Max Activating Token Index: 9

Text #4

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0870. Min Act: -0.0000

Data Index: 406558 (C4 (Web Text))

Max Activating Token Index: 246

Text #5

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.2129. Min Act: -0.0000

Data Index: 323297 (C4 (Web Text))

Max Activating Token Index: 428

Text #6

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.1170. Min Act: -0.0000

Data Index: 597859 (C4 (Web Text))

Max Activating Token Index: 442

Text #7

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0011. Min Act: -0.0000

Data Index: 1464710 (Python Code)

Max Activating Token Index: 645

Text #8

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.1256. Min Act: -0.0000

Data Index: 1434654 (Python Code)

Max Activating Token Index: 313

Text #9

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0636. Min Act: -0.0000

Data Index: 1544232 (Python Code)

Max Activating Token Index: 648

Text #10

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0801. Min Act: -0.0000

Data Index: 1459718 (Python Code)

Max Activating Token Index: 717

Text #11

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0407. Min Act: -0.0000

Data Index: 924229 (C4 (Web Text))

Max Activating Token Index: 857

Text #12

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.1030. Min Act: -0.0000

Data Index: 158479 (C4 (Web Text))

Max Activating Token Index: 307

Text #13

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0284. Min Act: -0.0000

Data Index: 1506174 (Python Code)

Max Activating Token Index: 781

Text #14

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0233. Min Act: -0.0000

Data Index: 1495662 (Python Code)

Max Activating Token Index: 709

Text #15

Max Range: 1.3517. Min Range: -1.3517

Max Act: 0.9196. Min Act: -0.0000

Data Index: 866728 (C4 (Web Text))

Max Activating Token Index: 623

Text #16

Max Range: 1.3517. Min Range: -1.3517

Max Act: 0.9969. Min Act: -0.0000

Data Index: 552272 (C4 (Web Text))

Max Activating Token Index: 381

Text #17

Max Range: 1.3517. Min Range: -1.3517

Max Act: 0.9824. Min Act: -0.0000

Data Index: 1483334 (Python Code)

Max Activating Token Index: 149

Text #18

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0687. Min Act: -0.0000

Data Index: 1067024 (C4 (Web Text))

Max Activating Token Index: 327

Text #19

Max Range: 1.3517. Min Range: -1.3517

Max Act: 1.0039. Min Act: -0.0000

Data Index: 1434691 (Python Code)

Max Activating Token Index: 751