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

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

Neuron 3535 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.3576. Min Act: -0.0001

Data Index: 165294 (C4 (Web Text))

Max Activating Token Index: 10

Text #1

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.3263. Min Act: -0.0001

Data Index: 924357 (C4 (Web Text))

Max Activating Token Index: 919

Text #2

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2519. Min Act: -0.0001

Data Index: 1612019 (Python Code)

Max Activating Token Index: 162

Text #3

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.3058. Min Act: -0.0001

Data Index: 246207 (C4 (Web Text))

Max Activating Token Index: 552

Text #4

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2691. Min Act: -0.0001

Data Index: 235309 (C4 (Web Text))

Max Activating Token Index: 876

Text #5

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2791. Min Act: -0.0001

Data Index: 1365317 (Python Code)

Max Activating Token Index: 323

Text #6

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2769. Min Act: -0.0001

Data Index: 1179068 (C4 (Web Text))

Max Activating Token Index: 129

Text #7

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2751. Min Act: -0.0001

Data Index: 1582205 (Python Code)

Max Activating Token Index: 595

Text #8

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2731. Min Act: -0.0001

Data Index: 2553 (C4 (Web Text))

Max Activating Token Index: 843

Text #9

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2696. Min Act: -0.0001

Data Index: 1076094 (C4 (Web Text))

Max Activating Token Index: 230

Text #10

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2618. Min Act: -0.0001

Data Index: 1624208 (Python Code)

Max Activating Token Index: 918

Text #11

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2468. Min Act: -0.0001

Data Index: 1418492 (Python Code)

Max Activating Token Index: 290

Text #12

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2574. Min Act: -0.0001

Data Index: 1545030 (Python Code)

Max Activating Token Index: 387

Text #13

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2555. Min Act: -0.0001

Data Index: 1651438 (Python Code)

Max Activating Token Index: 777

Text #14

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2554. Min Act: -0.0001

Data Index: 175062 (C4 (Web Text))

Max Activating Token Index: 269

Text #15

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2360. Min Act: -0.0001

Data Index: 427626 (C4 (Web Text))

Max Activating Token Index: 856

Text #16

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2899. Min Act: -0.0001

Data Index: 227437 (C4 (Web Text))

Max Activating Token Index: 723

Text #17

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2392. Min Act: -0.0001

Data Index: 1483676 (Python Code)

Max Activating Token Index: 326

Text #18

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.2416. Min Act: -0.0001

Data Index: 636719 (C4 (Web Text))

Max Activating Token Index: 678

Text #19

Max Range: 0.3576. Min Range: -0.3576

Max Act: 0.1923. Min Act: -0.0001

Data Index: 147073 (C4 (Web Text))

Max Activating Token Index: 283