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

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

Neuron 460 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6698. Min Act: -0.0000

Data Index: 44829 (C4 (Web Text))

Max Activating Token Index: 987

Text #1

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6433. Min Act: -0.0000

Data Index: 747237 (C4 (Web Text))

Max Activating Token Index: 796

Text #2

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6009. Min Act: -0.0000

Data Index: 228926 (C4 (Web Text))

Max Activating Token Index: 128

Text #3

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5934. Min Act: -0.0000

Data Index: 524859 (C4 (Web Text))

Max Activating Token Index: 830

Text #4

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6393. Min Act: -0.0000

Data Index: 1420923 (Python Code)

Max Activating Token Index: 717

Text #5

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5825. Min Act: -0.0000

Data Index: 161411 (C4 (Web Text))

Max Activating Token Index: 604

Text #6

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6085. Min Act: -0.0000

Data Index: 1029250 (C4 (Web Text))

Max Activating Token Index: 628

Text #7

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6068. Min Act: -0.0000

Data Index: 189969 (C4 (Web Text))

Max Activating Token Index: 851

Text #8

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6029. Min Act: -0.0000

Data Index: 131207 (C4 (Web Text))

Max Activating Token Index: 840

Text #9

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.6061. Min Act: -0.0000

Data Index: 1429784 (Python Code)

Max Activating Token Index: 621

Text #10

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5660. Min Act: -0.0000

Data Index: 850874 (C4 (Web Text))

Max Activating Token Index: 599

Text #11

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5607. Min Act: -0.0000

Data Index: 1278743 (C4 (Web Text))

Max Activating Token Index: 95

Text #12

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5712. Min Act: -0.0000

Data Index: 1233467 (C4 (Web Text))

Max Activating Token Index: 370

Text #13

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5651. Min Act: -0.0000

Data Index: 900037 (C4 (Web Text))

Max Activating Token Index: 548

Text #14

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5622. Min Act: -0.0000

Data Index: 1554056 (Python Code)

Max Activating Token Index: 701

Text #15

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5550. Min Act: -0.0000

Data Index: 1429275 (Python Code)

Max Activating Token Index: 57

Text #16

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5150. Min Act: -0.0000

Data Index: 1117934 (C4 (Web Text))

Max Activating Token Index: 186

Text #17

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5230. Min Act: -0.0000

Data Index: 1426790 (Python Code)

Max Activating Token Index: 587

Text #18

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5467. Min Act: -0.0000

Data Index: 1576990 (Python Code)

Max Activating Token Index: 459

Text #19

Max Range: 0.6698. Min Range: -0.6698

Max Act: 0.5490. Min Act: -0.0000

Data Index: 172169 (C4 (Web Text))

Max Activating Token Index: 750