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

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

Neuron 5653 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.4188. Min Range: -1.4188

Max Act: 1.4188. Min Act: -0.0001

Data Index: 1591448 (Python Code)

Max Activating Token Index: 3

Text #1

Max Range: 1.4188. Min Range: -1.4188

Max Act: 1.1761. Min Act: -0.0001

Data Index: 1605900 (Python Code)

Max Activating Token Index: 5

Text #2

Max Range: 1.4188. Min Range: -1.4188

Max Act: 1.0317. Min Act: -0.0001

Data Index: 1587475 (Python Code)

Max Activating Token Index: 65

Text #3

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.9314. Min Act: -0.0001

Data Index: 1432332 (Python Code)

Max Activating Token Index: 6

Text #4

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8628. Min Act: -0.0001

Data Index: 1574483 (Python Code)

Max Activating Token Index: 19

Text #5

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8770. Min Act: -0.0001

Data Index: 392452 (C4 (Web Text))

Max Activating Token Index: 111

Text #6

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.9049. Min Act: -0.0001

Data Index: 1593156 (Python Code)

Max Activating Token Index: 5

Text #7

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8041. Min Act: -0.0001

Data Index: 1597209 (Python Code)

Max Activating Token Index: 390

Text #8

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8816. Min Act: -0.0001

Data Index: 1520593 (Python Code)

Max Activating Token Index: 4

Text #9

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8816. Min Act: -0.0001

Data Index: 1404694 (Python Code)

Max Activating Token Index: 4

Text #10

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8763. Min Act: -0.0001

Data Index: 1586821 (Python Code)

Max Activating Token Index: 6

Text #11

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8616. Min Act: -0.0001

Data Index: 1495906 (Python Code)

Max Activating Token Index: 19

Text #12

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8604. Min Act: -0.0001

Data Index: 1472243 (Python Code)

Max Activating Token Index: 38

Text #13

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8292. Min Act: -0.0001

Data Index: 1450214 (Python Code)

Max Activating Token Index: 5

Text #14

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8513. Min Act: -0.0001

Data Index: 1475488 (Python Code)

Max Activating Token Index: 80

Text #15

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8155. Min Act: -0.0001

Data Index: 1445663 (Python Code)

Max Activating Token Index: 22

Text #16

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8520. Min Act: -0.0001

Data Index: 250382 (C4 (Web Text))

Max Activating Token Index: 742

Text #17

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8560. Min Act: -0.0001

Data Index: 1392828 (Python Code)

Max Activating Token Index: 374

Text #18

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.7503. Min Act: -0.0001

Data Index: 807151 (C4 (Web Text))

Max Activating Token Index: 787

Text #19

Max Range: 1.4188. Min Range: -1.4188

Max Act: 0.8584. Min Act: -0.0001

Data Index: 1123524 (C4 (Web Text))

Max Activating Token Index: 924