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

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

Neuron 4640 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.8287. Min Act: -0.0000

Data Index: 1533396 (Python Code)

Max Activating Token Index: 719

Text #1

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.7049. Min Act: -0.0000

Data Index: 1626711 (Python Code)

Max Activating Token Index: 1001

Text #2

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.7108. Min Act: -0.0000

Data Index: 1560817 (Python Code)

Max Activating Token Index: 564

Text #3

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6562. Min Act: -0.0000

Data Index: 149716 (C4 (Web Text))

Max Activating Token Index: 666

Text #4

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6569. Min Act: 0.0001

Data Index: 1161373 (C4 (Web Text))

Max Activating Token Index: 92

Text #5

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6569. Min Act: -0.0000

Data Index: 231601 (C4 (Web Text))

Max Activating Token Index: 423

Text #6

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6207. Min Act: -0.0000

Data Index: 410441 (C4 (Web Text))

Max Activating Token Index: 98

Text #7

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6034. Min Act: -0.0000

Data Index: 753608 (C4 (Web Text))

Max Activating Token Index: 444

Text #8

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6212. Min Act: -0.0000

Data Index: 638587 (C4 (Web Text))

Max Activating Token Index: 21

Text #9

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5927. Min Act: -0.0000

Data Index: 623875 (C4 (Web Text))

Max Activating Token Index: 510

Text #10

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6771. Min Act: -0.0000

Data Index: 1657100 (Python Code)

Max Activating Token Index: 152

Text #11

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6029. Min Act: -0.0000

Data Index: 1454339 (Python Code)

Max Activating Token Index: 54

Text #12

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6218. Min Act: -0.0000

Data Index: 223714 (C4 (Web Text))

Max Activating Token Index: 767

Text #13

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.6013. Min Act: -0.0000

Data Index: 1574478 (Python Code)

Max Activating Token Index: 431

Text #14

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5980. Min Act: 0.0000

Data Index: 160358 (C4 (Web Text))

Max Activating Token Index: 985

Text #15

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5964. Min Act: -0.0000

Data Index: 1488821 (Python Code)

Max Activating Token Index: 74

Text #16

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5232. Min Act: -0.0000

Data Index: 1451189 (Python Code)

Max Activating Token Index: 830

Text #17

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5691. Min Act: 0.0000

Data Index: 1246927 (C4 (Web Text))

Max Activating Token Index: 320

Text #18

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5783. Min Act: -0.0000

Data Index: 1624000 (Python Code)

Max Activating Token Index: 168

Text #19

Max Range: 0.8287. Min Range: -0.8287

Max Act: 0.5560. Min Act: -0.0000

Data Index: 5930 (C4 (Web Text))

Max Activating Token Index: 885