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

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

Neuron 4724 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1523. Min Act: -0.0000

Data Index: 724101 (C4 (Web Text))

Max Activating Token Index: 351

Text #1

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1343. Min Act: -0.0000

Data Index: 653090 (C4 (Web Text))

Max Activating Token Index: 117

Text #2

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1277. Min Act: -0.0000

Data Index: 634524 (C4 (Web Text))

Max Activating Token Index: 949

Text #3

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1368. Min Act: -0.0000

Data Index: 28052 (C4 (Web Text))

Max Activating Token Index: 381

Text #4

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1283. Min Act: -0.0000

Data Index: 212468 (C4 (Web Text))

Max Activating Token Index: 625

Text #5

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1289. Min Act: -0.0000

Data Index: 774352 (C4 (Web Text))

Max Activating Token Index: 584

Text #6

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1359. Min Act: -0.0000

Data Index: 430310 (C4 (Web Text))

Max Activating Token Index: 693

Text #7

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1337. Min Act: -0.0000

Data Index: 21999 (C4 (Web Text))

Max Activating Token Index: 610

Text #8

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1163. Min Act: -0.0000

Data Index: 1539298 (Python Code)

Max Activating Token Index: 930

Text #9

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1208. Min Act: -0.0000

Data Index: 1468355 (Python Code)

Max Activating Token Index: 512

Text #10

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1211. Min Act: -0.0000

Data Index: 1630073 (Python Code)

Max Activating Token Index: 771

Text #11

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1250. Min Act: -0.0000

Data Index: 1134256 (C4 (Web Text))

Max Activating Token Index: 794

Text #12

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1188. Min Act: -0.0000

Data Index: 626369 (C4 (Web Text))

Max Activating Token Index: 371

Text #13

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1253. Min Act: -0.0000

Data Index: 1049378 (C4 (Web Text))

Max Activating Token Index: 323

Text #14

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1152. Min Act: -0.0000

Data Index: 975641 (C4 (Web Text))

Max Activating Token Index: 387

Text #15

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1234. Min Act: -0.0000

Data Index: 1397470 (Python Code)

Max Activating Token Index: 899

Text #16

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1113. Min Act: -0.0000

Data Index: 1222535 (C4 (Web Text))

Max Activating Token Index: 994

Text #17

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1234. Min Act: -0.0000

Data Index: 171376 (C4 (Web Text))

Max Activating Token Index: 703

Text #18

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1105. Min Act: -0.0000

Data Index: 1520684 (Python Code)

Max Activating Token Index: 605

Text #19

Max Range: 0.1523. Min Range: -0.1523

Max Act: 0.1058. Min Act: -0.0000

Data Index: 439618 (C4 (Web Text))

Max Activating Token Index: 620