Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 86 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2286. Min Act: -0.0001

Data Index: 1285357 (C4 (Web Text))

Max Activating Token Index: 256

Text #1

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2283. Min Act: -0.0001

Data Index: 1478961 (Python Code)

Max Activating Token Index: 821

Text #2

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2344. Min Act: -0.0001

Data Index: 482422 (C4 (Web Text))

Max Activating Token Index: 929

Text #3

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2235. Min Act: -0.0001

Data Index: 861244 (C4 (Web Text))

Max Activating Token Index: 332

Text #4

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2080. Min Act: -0.0001

Data Index: 1571833 (Python Code)

Max Activating Token Index: 939

Text #5

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2309. Min Act: -0.0001

Data Index: 592468 (C4 (Web Text))

Max Activating Token Index: 291

Text #6

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2048. Min Act: -0.0001

Data Index: 1280729 (C4 (Web Text))

Max Activating Token Index: 168

Text #7

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1991. Min Act: -0.0001

Data Index: 518657 (C4 (Web Text))

Max Activating Token Index: 875

Text #8

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.2046. Min Act: -0.0001

Data Index: 1540350 (Python Code)

Max Activating Token Index: 920

Text #9

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1971. Min Act: -0.0001

Data Index: 1608493 (Python Code)

Max Activating Token Index: 231

Text #10

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1896. Min Act: -0.0001

Data Index: 8102 (C4 (Web Text))

Max Activating Token Index: 969

Text #11

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1948. Min Act: -0.0001

Data Index: 898096 (C4 (Web Text))

Max Activating Token Index: 168

Text #12

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1928. Min Act: -0.0001

Data Index: 747369 (C4 (Web Text))

Max Activating Token Index: 134

Text #13

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1912. Min Act: -0.0001

Data Index: 1083379 (C4 (Web Text))

Max Activating Token Index: 850

Text #14

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1824. Min Act: -0.0001

Data Index: 1527017 (Python Code)

Max Activating Token Index: 177

Text #15

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1888. Min Act: -0.0001

Data Index: 972964 (C4 (Web Text))

Max Activating Token Index: 802

Text #16

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1884. Min Act: -0.0001

Data Index: 359677 (C4 (Web Text))

Max Activating Token Index: 472

Text #17

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1936. Min Act: -0.0001

Data Index: 1205688 (C4 (Web Text))

Max Activating Token Index: 6

Text #18

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1944. Min Act: -0.0001

Data Index: 459794 (C4 (Web Text))

Max Activating Token Index: 995

Text #19

Max Range: 0.2344. Min Range: -0.2344

Max Act: 0.1839. Min Act: -0.0001

Data Index: 4852 (C4 (Web Text))

Max Activating Token Index: 177