Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2780 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1889. Min Act: -0.0001

Data Index: 567206 (C4 (Web Text))

Max Activating Token Index: 866

Text #1

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1654. Min Act: -0.0001

Data Index: 284645 (C4 (Web Text))

Max Activating Token Index: 328

Text #2

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1664. Min Act: -0.0001

Data Index: 254295 (C4 (Web Text))

Max Activating Token Index: 580

Text #3

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1755. Min Act: -0.0001

Data Index: 951939 (C4 (Web Text))

Max Activating Token Index: 749

Text #4

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1563. Min Act: -0.0001

Data Index: 1223471 (C4 (Web Text))

Max Activating Token Index: 743

Text #5

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1645. Min Act: -0.0001

Data Index: 1261081 (C4 (Web Text))

Max Activating Token Index: 383

Text #6

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1549. Min Act: -0.0001

Data Index: 952999 (C4 (Web Text))

Max Activating Token Index: 389

Text #7

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1502. Min Act: -0.0001

Data Index: 1233418 (C4 (Web Text))

Max Activating Token Index: 538

Text #8

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1503. Min Act: -0.0001

Data Index: 449486 (C4 (Web Text))

Max Activating Token Index: 215

Text #9

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1481. Min Act: -0.0001

Data Index: 576352 (C4 (Web Text))

Max Activating Token Index: 312

Text #10

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1511. Min Act: -0.0001

Data Index: 1273685 (C4 (Web Text))

Max Activating Token Index: 130

Text #11

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1448. Min Act: -0.0001

Data Index: 921694 (C4 (Web Text))

Max Activating Token Index: 400

Text #12

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1297. Min Act: -0.0001

Data Index: 476884 (C4 (Web Text))

Max Activating Token Index: 743

Text #13

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1334. Min Act: -0.0001

Data Index: 1074652 (C4 (Web Text))

Max Activating Token Index: 834

Text #14

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1473. Min Act: -0.0001

Data Index: 770393 (C4 (Web Text))

Max Activating Token Index: 736

Text #15

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1319. Min Act: -0.0001

Data Index: 660759 (C4 (Web Text))

Max Activating Token Index: 584

Text #16

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1391. Min Act: -0.0001

Data Index: 1508625 (Python Code)

Max Activating Token Index: 672

Text #17

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1290. Min Act: -0.0001

Data Index: 932105 (C4 (Web Text))

Max Activating Token Index: 1008

Text #18

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1286. Min Act: -0.0001

Data Index: 1118145 (C4 (Web Text))

Max Activating Token Index: 841

Text #19

Max Range: 0.1889. Min Range: -0.1889

Max Act: 0.1228. Min Act: -0.0001

Data Index: 187762 (C4 (Web Text))

Max Activating Token Index: 755