Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 4000 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1616. Min Act: -0.0000

Data Index: 878641 (C4 (Web Text))

Max Activating Token Index: 253

Text #1

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1657. Min Act: -0.0000

Data Index: 92309 (C4 (Web Text))

Max Activating Token Index: 466

Text #2

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1507. Min Act: -0.0000

Data Index: 937031 (C4 (Web Text))

Max Activating Token Index: 280

Text #3

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1546. Min Act: -0.0000

Data Index: 1583593 (Python Code)

Max Activating Token Index: 240

Text #4

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1517. Min Act: -0.0000

Data Index: 551303 (C4 (Web Text))

Max Activating Token Index: 490

Text #5

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1473. Min Act: -0.0000

Data Index: 953102 (C4 (Web Text))

Max Activating Token Index: 695

Text #6

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1504. Min Act: -0.0000

Data Index: 752607 (C4 (Web Text))

Max Activating Token Index: 852

Text #7

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1454. Min Act: -0.0000

Data Index: 1318707 (C4 (Web Text))

Max Activating Token Index: 100

Text #8

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1507. Min Act: -0.0000

Data Index: 870563 (C4 (Web Text))

Max Activating Token Index: 611

Text #9

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1421. Min Act: -0.0000

Data Index: 231464 (C4 (Web Text))

Max Activating Token Index: 108

Text #10

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1412. Min Act: -0.0000

Data Index: 1620546 (Python Code)

Max Activating Token Index: 873

Text #11

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1450. Min Act: -0.0000

Data Index: 149598 (C4 (Web Text))

Max Activating Token Index: 140

Text #12

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1406. Min Act: -0.0000

Data Index: 845985 (C4 (Web Text))

Max Activating Token Index: 215

Text #13

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1252. Min Act: -0.0000

Data Index: 194179 (C4 (Web Text))

Max Activating Token Index: 675

Text #14

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1338. Min Act: -0.0000

Data Index: 1145392 (C4 (Web Text))

Max Activating Token Index: 114

Text #15

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1330. Min Act: -0.0000

Data Index: 62851 (C4 (Web Text))

Max Activating Token Index: 482

Text #16

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1322. Min Act: -0.0000

Data Index: 719738 (C4 (Web Text))

Max Activating Token Index: 627

Text #17

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1322. Min Act: -0.0000

Data Index: 1390762 (Python Code)

Max Activating Token Index: 757

Text #18

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1303. Min Act: -0.0000

Data Index: 469980 (C4 (Web Text))

Max Activating Token Index: 486

Text #19

Max Range: 0.1657. Min Range: -0.1657

Max Act: 0.1352. Min Act: -0.0000

Data Index: 419481 (C4 (Web Text))

Max Activating Token Index: 107