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

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

Neuron 683 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8450. Min Act: -0.0001

Data Index: 795115 (C4 (Web Text))

Max Activating Token Index: 201

Text #1

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.9045. Min Act: -0.0001

Data Index: 1322520 (C4 (Web Text))

Max Activating Token Index: 326

Text #2

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8936. Min Act: -0.0001

Data Index: 204340 (C4 (Web Text))

Max Activating Token Index: 577

Text #3

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.9167. Min Act: -0.0001

Data Index: 40847 (C4 (Web Text))

Max Activating Token Index: 63

Text #4

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8612. Min Act: -0.0001

Data Index: 226589 (C4 (Web Text))

Max Activating Token Index: 809

Text #5

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8796. Min Act: -0.0001

Data Index: 1601035 (Python Code)

Max Activating Token Index: 325

Text #6

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8773. Min Act: -0.0001

Data Index: 421607 (C4 (Web Text))

Max Activating Token Index: 960

Text #7

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.9373. Min Act: -0.0001

Data Index: 1370375 (Python Code)

Max Activating Token Index: 255

Text #8

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8627. Min Act: -0.0001

Data Index: 1090839 (C4 (Web Text))

Max Activating Token Index: 693

Text #9

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8540. Min Act: -0.0001

Data Index: 293510 (C4 (Web Text))

Max Activating Token Index: 861

Text #10

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8555. Min Act: -0.0001

Data Index: 1460314 (Python Code)

Max Activating Token Index: 39

Text #11

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8615. Min Act: -0.0001

Data Index: 1629601 (Python Code)

Max Activating Token Index: 884

Text #12

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8675. Min Act: -0.0001

Data Index: 1510672 (Python Code)

Max Activating Token Index: 180

Text #13

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8427. Min Act: -0.0001

Data Index: 1532774 (Python Code)

Max Activating Token Index: 577

Text #14

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8355. Min Act: -0.0001

Data Index: 156255 (C4 (Web Text))

Max Activating Token Index: 699

Text #15

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8271. Min Act: -0.0001

Data Index: 1647823 (Python Code)

Max Activating Token Index: 375

Text #16

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8993. Min Act: -0.0001

Data Index: 1577425 (Python Code)

Max Activating Token Index: 243

Text #17

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.8408. Min Act: -0.0001

Data Index: 424006 (C4 (Web Text))

Max Activating Token Index: 31

Text #18

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.7996. Min Act: -0.0001

Data Index: 582675 (C4 (Web Text))

Max Activating Token Index: 600

Text #19

Max Range: 0.9373. Min Range: -0.9373

Max Act: 0.7735. Min Act: -0.0001

Data Index: 915400 (C4 (Web Text))

Max Activating Token Index: 808