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

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

Neuron 2904 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.4192. Min Range: -5.4192

Max Act: 5.4192. Min Act: -0.0000

Data Index: 1559545 (Python Code)

Max Activating Token Index: 584

Text #1

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.8618. Min Act: -0.0000

Data Index: 1585648 (Python Code)

Max Activating Token Index: 673

Text #2

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.8485. Min Act: -0.0000

Data Index: 615365 (C4 (Web Text))

Max Activating Token Index: 113

Text #3

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.6610. Min Act: -0.0000

Data Index: 791022 (C4 (Web Text))

Max Activating Token Index: 462

Text #4

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.2050. Min Act: -0.0000

Data Index: 1612564 (Python Code)

Max Activating Token Index: 95

Text #5

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.1394. Min Act: -0.0000

Data Index: 1199835 (C4 (Web Text))

Max Activating Token Index: 681

Text #6

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.1035. Min Act: -0.0000

Data Index: 1331833 (C4 (Web Text))

Max Activating Token Index: 70

Text #7

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.4459. Min Act: -0.0000

Data Index: 438838 (C4 (Web Text))

Max Activating Token Index: 95

Text #8

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.0076. Min Act: -0.0000

Data Index: 7521 (C4 (Web Text))

Max Activating Token Index: 243

Text #9

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.6900. Min Act: -0.0000

Data Index: 1390412 (Python Code)

Max Activating Token Index: 136

Text #10

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.8944. Min Act: -0.0000

Data Index: 838208 (C4 (Web Text))

Max Activating Token Index: 632

Text #11

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.3386. Min Act: -0.0000

Data Index: 817446 (C4 (Web Text))

Max Activating Token Index: 464

Text #12

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.6680. Min Act: -0.0000

Data Index: 1638703 (Python Code)

Max Activating Token Index: 305

Text #13

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.2817. Min Act: -0.0000

Data Index: 13130 (C4 (Web Text))

Max Activating Token Index: 248

Text #14

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.8681. Min Act: -0.0000

Data Index: 360192 (C4 (Web Text))

Max Activating Token Index: 621

Text #15

Max Range: 5.4192. Min Range: -5.4192

Max Act: 4.0125. Min Act: -0.0000

Data Index: 705250 (C4 (Web Text))

Max Activating Token Index: 728

Text #16

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.9425. Min Act: -0.0000

Data Index: 108287 (C4 (Web Text))

Max Activating Token Index: 582

Text #17

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.5850. Min Act: -0.0000

Data Index: 1056657 (C4 (Web Text))

Max Activating Token Index: 519

Text #18

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.8311. Min Act: -0.0000

Data Index: 1576686 (Python Code)

Max Activating Token Index: 659

Text #19

Max Range: 5.4192. Min Range: -5.4192

Max Act: 3.6690. Min Act: -0.0000

Data Index: 724855 (C4 (Web Text))

Max Activating Token Index: 629