Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1719 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2924. Min Act: -0.0002

Data Index: 109209 (C4 (Web Text))

Max Activating Token Index: 239

Text #1

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2440. Min Act: -0.0002

Data Index: 760498 (C4 (Web Text))

Max Activating Token Index: 133

Text #2

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2283. Min Act: -0.0002

Data Index: 1250391 (C4 (Web Text))

Max Activating Token Index: 540

Text #3

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2287. Min Act: -0.0002

Data Index: 576928 (C4 (Web Text))

Max Activating Token Index: 39

Text #4

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2250. Min Act: -0.0002

Data Index: 531798 (C4 (Web Text))

Max Activating Token Index: 108

Text #5

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2316. Min Act: -0.0002

Data Index: 409944 (C4 (Web Text))

Max Activating Token Index: 555

Text #6

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2186. Min Act: -0.0002

Data Index: 222573 (C4 (Web Text))

Max Activating Token Index: 47

Text #7

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2170. Min Act: -0.0002

Data Index: 962944 (C4 (Web Text))

Max Activating Token Index: 646

Text #8

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2193. Min Act: -0.0002

Data Index: 1515521 (Python Code)

Max Activating Token Index: 892

Text #9

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2152. Min Act: -0.0002

Data Index: 1646200 (Python Code)

Max Activating Token Index: 589

Text #10

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2111. Min Act: -0.0002

Data Index: 900147 (C4 (Web Text))

Max Activating Token Index: 787

Text #11

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2149. Min Act: -0.0002

Data Index: 1050274 (C4 (Web Text))

Max Activating Token Index: 751

Text #12

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2064. Min Act: -0.0002

Data Index: 1390138 (Python Code)

Max Activating Token Index: 543

Text #13

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.1992. Min Act: -0.0002

Data Index: 739370 (C4 (Web Text))

Max Activating Token Index: 969

Text #14

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2086. Min Act: -0.0002

Data Index: 107694 (C4 (Web Text))

Max Activating Token Index: 74

Text #15

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2008. Min Act: -0.0002

Data Index: 116149 (C4 (Web Text))

Max Activating Token Index: 388

Text #16

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.1998. Min Act: -0.0002

Data Index: 1608920 (Python Code)

Max Activating Token Index: 123

Text #17

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2068. Min Act: -0.0002

Data Index: 1532608 (Python Code)

Max Activating Token Index: 654

Text #18

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.2016. Min Act: -0.0002

Data Index: 619811 (C4 (Web Text))

Max Activating Token Index: 141

Text #19

Max Range: 0.2924. Min Range: -0.2924

Max Act: 0.1891. Min Act: -0.0002

Data Index: 843130 (C4 (Web Text))

Max Activating Token Index: 95