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

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

Neuron 1928 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6724. Min Act: -0.0001

Data Index: 601823 (C4 (Web Text))

Max Activating Token Index: 871

Text #1

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.7284. Min Act: -0.0001

Data Index: 133220 (C4 (Web Text))

Max Activating Token Index: 719

Text #2

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6606. Min Act: -0.0001

Data Index: 545780 (C4 (Web Text))

Max Activating Token Index: 866

Text #3

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6451. Min Act: -0.0001

Data Index: 1430959 (Python Code)

Max Activating Token Index: 125

Text #4

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6152. Min Act: -0.0001

Data Index: 1394252 (Python Code)

Max Activating Token Index: 45

Text #5

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6457. Min Act: -0.0001

Data Index: 235227 (C4 (Web Text))

Max Activating Token Index: 938

Text #6

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6588. Min Act: -0.0001

Data Index: 195539 (C4 (Web Text))

Max Activating Token Index: 951

Text #7

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6100. Min Act: -0.0001

Data Index: 1024725 (C4 (Web Text))

Max Activating Token Index: 1016

Text #8

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6430. Min Act: -0.0001

Data Index: 406360 (C4 (Web Text))

Max Activating Token Index: 723

Text #9

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6146. Min Act: -0.0001

Data Index: 1443933 (Python Code)

Max Activating Token Index: 494

Text #10

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5556. Min Act: -0.0000

Data Index: 1452438 (Python Code)

Max Activating Token Index: 196

Text #11

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.6223. Min Act: -0.0001

Data Index: 1565994 (Python Code)

Max Activating Token Index: 575

Text #12

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5896. Min Act: -0.0001

Data Index: 1131560 (C4 (Web Text))

Max Activating Token Index: 908

Text #13

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5471. Min Act: -0.0001

Data Index: 1256997 (C4 (Web Text))

Max Activating Token Index: 792

Text #14

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5662. Min Act: -0.0001

Data Index: 612939 (C4 (Web Text))

Max Activating Token Index: 992

Text #15

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5675. Min Act: -0.0001

Data Index: 1507124 (Python Code)

Max Activating Token Index: 935

Text #16

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5417. Min Act: -0.0001

Data Index: 639960 (C4 (Web Text))

Max Activating Token Index: 115

Text #17

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5402. Min Act: -0.0001

Data Index: 1558968 (Python Code)

Max Activating Token Index: 1016

Text #18

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5006. Min Act: -0.0001

Data Index: 774142 (C4 (Web Text))

Max Activating Token Index: 738

Text #19

Max Range: 0.7284. Min Range: -0.7284

Max Act: 0.5469. Min Act: -0.0001

Data Index: 341744 (C4 (Web Text))

Max Activating Token Index: 80