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

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

Neuron 931 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6611. Min Act: -0.0001

Data Index: 1564146 (Python Code)

Max Activating Token Index: 856

Text #1

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6310. Min Act: -0.0001

Data Index: 670401 (C4 (Web Text))

Max Activating Token Index: 227

Text #2

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6276. Min Act: -0.0001

Data Index: 479431 (C4 (Web Text))

Max Activating Token Index: 282

Text #3

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5644. Min Act: -0.0001

Data Index: 688097 (C4 (Web Text))

Max Activating Token Index: 180

Text #4

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6133. Min Act: -0.0001

Data Index: 1536604 (Python Code)

Max Activating Token Index: 830

Text #5

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6369. Min Act: -0.0001

Data Index: 622777 (C4 (Web Text))

Max Activating Token Index: 305

Text #6

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5506. Min Act: -0.0001

Data Index: 1630187 (Python Code)

Max Activating Token Index: 983

Text #7

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.4907. Min Act: -0.0001

Data Index: 1478143 (Python Code)

Max Activating Token Index: 632

Text #8

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.6267. Min Act: -0.0001

Data Index: 1561663 (Python Code)

Max Activating Token Index: 900

Text #9

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5281. Min Act: -0.0001

Data Index: 1490030 (Python Code)

Max Activating Token Index: 450

Text #10

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5583. Min Act: -0.0001

Data Index: 187355 (C4 (Web Text))

Max Activating Token Index: 285

Text #11

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5263. Min Act: -0.0001

Data Index: 1480664 (Python Code)

Max Activating Token Index: 763

Text #12

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5092. Min Act: -0.0001

Data Index: 750670 (C4 (Web Text))

Max Activating Token Index: 794

Text #13

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5435. Min Act: -0.0001

Data Index: 1443524 (Python Code)

Max Activating Token Index: 437

Text #14

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.4912. Min Act: -0.0001

Data Index: 1359981 (Python Code)

Max Activating Token Index: 63

Text #15

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5317. Min Act: -0.0001

Data Index: 1453199 (Python Code)

Max Activating Token Index: 65

Text #16

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5440. Min Act: -0.0001

Data Index: 1541622 (Python Code)

Max Activating Token Index: 469

Text #17

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.5061. Min Act: -0.0001

Data Index: 1161761 (C4 (Web Text))

Max Activating Token Index: 178

Text #18

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.4721. Min Act: -0.0001

Data Index: 1654868 (Python Code)

Max Activating Token Index: 33

Text #19

Max Range: 0.6611. Min Range: -0.6611

Max Act: 0.4484. Min Act: -0.0001

Data Index: 1001735 (C4 (Web Text))

Max Activating Token Index: 95