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

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

Neuron 503 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: 0.1453. Min Range: -0.1453

Max Act: 0.1355. Min Act: -0.0002

Data Index: 348484 (C4 (Web Text))

Max Activating Token Index: 967

Text #1

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1374. Min Act: -0.0002

Data Index: 680264 (C4 (Web Text))

Max Activating Token Index: 572

Text #2

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1340. Min Act: -0.0002

Data Index: 1034763 (C4 (Web Text))

Max Activating Token Index: 971

Text #3

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1219. Min Act: -0.0002

Data Index: 354406 (C4 (Web Text))

Max Activating Token Index: 967

Text #4

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1140. Min Act: -0.0002

Data Index: 1237689 (C4 (Web Text))

Max Activating Token Index: 506

Text #5

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1381. Min Act: -0.0002

Data Index: 763511 (C4 (Web Text))

Max Activating Token Index: 734

Text #6

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1453. Min Act: -0.0002

Data Index: 1217552 (C4 (Web Text))

Max Activating Token Index: 755

Text #7

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1209. Min Act: -0.0002

Data Index: 1200763 (C4 (Web Text))

Max Activating Token Index: 996

Text #8

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1346. Min Act: -0.0002

Data Index: 976930 (C4 (Web Text))

Max Activating Token Index: 947

Text #9

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1311. Min Act: -0.0002

Data Index: 1344969 (C4 (Web Text))

Max Activating Token Index: 742

Text #10

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1278. Min Act: -0.0002

Data Index: 641764 (C4 (Web Text))

Max Activating Token Index: 672

Text #11

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1282. Min Act: -0.0002

Data Index: 43016 (C4 (Web Text))

Max Activating Token Index: 770

Text #12

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1322. Min Act: -0.0002

Data Index: 204422 (C4 (Web Text))

Max Activating Token Index: 638

Text #13

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1308. Min Act: -0.0002

Data Index: 1333584 (C4 (Web Text))

Max Activating Token Index: 740

Text #14

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1332. Min Act: -0.0002

Data Index: 1206949 (C4 (Web Text))

Max Activating Token Index: 777

Text #15

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1248. Min Act: -0.0002

Data Index: 617127 (C4 (Web Text))

Max Activating Token Index: 981

Text #16

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1148. Min Act: -0.0002

Data Index: 626234 (C4 (Web Text))

Max Activating Token Index: 648

Text #17

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1232. Min Act: -0.0002

Data Index: 463155 (C4 (Web Text))

Max Activating Token Index: 738

Text #18

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1235. Min Act: -0.0002

Data Index: 334679 (C4 (Web Text))

Max Activating Token Index: 194

Text #19

Max Range: 0.1453. Min Range: -0.1453

Max Act: 0.1149. Min Act: -0.0002

Data Index: 388497 (C4 (Web Text))

Max Activating Token Index: 412