Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 684 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.4367. Min Act: -0.0000

Data Index: 1424645 (Python Code)

Max Activating Token Index: 1006

Text #1

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.3087. Min Act: -0.0000

Data Index: 671059 (C4 (Web Text))

Max Activating Token Index: 708

Text #2

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.3057. Min Act: -0.0000

Data Index: 251171 (C4 (Web Text))

Max Activating Token Index: 60

Text #3

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2726. Min Act: -0.0000

Data Index: 768969 (C4 (Web Text))

Max Activating Token Index: 132

Text #4

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2397. Min Act: -0.0000

Data Index: 1435761 (Python Code)

Max Activating Token Index: 809

Text #5

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2291. Min Act: -0.0000

Data Index: 394560 (C4 (Web Text))

Max Activating Token Index: 690

Text #6

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2228. Min Act: -0.0000

Data Index: 263760 (C4 (Web Text))

Max Activating Token Index: 478

Text #7

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2380. Min Act: -0.0000

Data Index: 1407244 (Python Code)

Max Activating Token Index: 378

Text #8

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2390. Min Act: -0.0000

Data Index: 551203 (C4 (Web Text))

Max Activating Token Index: 678

Text #9

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2389. Min Act: -0.0000

Data Index: 1383338 (Python Code)

Max Activating Token Index: 817

Text #10

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2209. Min Act: -0.0000

Data Index: 1353218 (C4 (Web Text))

Max Activating Token Index: 414

Text #11

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1819. Min Act: -0.0000

Data Index: 1496553 (Python Code)

Max Activating Token Index: 187

Text #12

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1923. Min Act: -0.0000

Data Index: 337125 (C4 (Web Text))

Max Activating Token Index: 666

Text #13

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.2009. Min Act: -0.0000

Data Index: 1189613 (C4 (Web Text))

Max Activating Token Index: 733

Text #14

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1958. Min Act: -0.0000

Data Index: 1141208 (C4 (Web Text))

Max Activating Token Index: 461

Text #15

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1869. Min Act: -0.0000

Data Index: 1338314 (C4 (Web Text))

Max Activating Token Index: 907

Text #16

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1686. Min Act: -0.0000

Data Index: 1221733 (C4 (Web Text))

Max Activating Token Index: 465

Text #17

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1764. Min Act: -0.0000

Data Index: 15037 (C4 (Web Text))

Max Activating Token Index: 297

Text #18

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1854. Min Act: -0.0000

Data Index: 1091626 (C4 (Web Text))

Max Activating Token Index: 229

Text #19

Max Range: 0.4367. Min Range: -0.4367

Max Act: 0.1827. Min Act: -0.0000

Data Index: 157826 (C4 (Web Text))

Max Activating Token Index: 541