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

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

Neuron 3119 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1389. Min Act: -0.0001

Data Index: 177225 (C4 (Web Text))

Max Activating Token Index: 574

Text #1

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1283. Min Act: -0.0001

Data Index: 329348 (C4 (Web Text))

Max Activating Token Index: 662

Text #2

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1141. Min Act: -0.0001

Data Index: 1599941 (Python Code)

Max Activating Token Index: 575

Text #3

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1219. Min Act: -0.0001

Data Index: 1563028 (Python Code)

Max Activating Token Index: 175

Text #4

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1257. Min Act: -0.0001

Data Index: 67377 (C4 (Web Text))

Max Activating Token Index: 999

Text #5

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1199. Min Act: -0.0001

Data Index: 812255 (C4 (Web Text))

Max Activating Token Index: 26

Text #6

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1287. Min Act: -0.0001

Data Index: 1565737 (Python Code)

Max Activating Token Index: 963

Text #7

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1318. Min Act: -0.0001

Data Index: 429681 (C4 (Web Text))

Max Activating Token Index: 686

Text #8

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1196. Min Act: -0.0001

Data Index: 188098 (C4 (Web Text))

Max Activating Token Index: 868

Text #9

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1267. Min Act: -0.0001

Data Index: 1250778 (C4 (Web Text))

Max Activating Token Index: 125

Text #10

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1180. Min Act: -0.0001

Data Index: 1491155 (Python Code)

Max Activating Token Index: 448

Text #11

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1203. Min Act: -0.0001

Data Index: 770154 (C4 (Web Text))

Max Activating Token Index: 110

Text #12

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1195. Min Act: -0.0001

Data Index: 1570879 (Python Code)

Max Activating Token Index: 303

Text #13

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1174. Min Act: -0.0001

Data Index: 386825 (C4 (Web Text))

Max Activating Token Index: 107

Text #14

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1165. Min Act: -0.0001

Data Index: 1004527 (C4 (Web Text))

Max Activating Token Index: 84

Text #15

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1225. Min Act: -0.0001

Data Index: 227669 (C4 (Web Text))

Max Activating Token Index: 756

Text #16

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1167. Min Act: -0.0001

Data Index: 1656587 (Python Code)

Max Activating Token Index: 781

Text #17

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1192. Min Act: -0.0001

Data Index: 1432408 (Python Code)

Max Activating Token Index: 845

Text #18

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1106. Min Act: -0.0001

Data Index: 1232112 (C4 (Web Text))

Max Activating Token Index: 179

Text #19

Max Range: 0.1389. Min Range: -0.1389

Max Act: 0.1073. Min Act: -0.0001

Data Index: 1324050 (C4 (Web Text))

Max Activating Token Index: 825