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

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

Neuron 2589 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: 0.3099. Min Range: -0.3099

Max Act: 0.3081. Min Act: -0.0000

Data Index: 682714 (C4 (Web Text))

Max Activating Token Index: 377

Text #1

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2970. Min Act: -0.0000

Data Index: 789957 (C4 (Web Text))

Max Activating Token Index: 279

Text #2

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.3099. Min Act: -0.0000

Data Index: 602122 (C4 (Web Text))

Max Activating Token Index: 679

Text #3

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2958. Min Act: -0.0000

Data Index: 771738 (C4 (Web Text))

Max Activating Token Index: 385

Text #4

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2871. Min Act: -0.0000

Data Index: 1398750 (Python Code)

Max Activating Token Index: 480

Text #5

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2739. Min Act: -0.0000

Data Index: 171431 (C4 (Web Text))

Max Activating Token Index: 197

Text #6

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2512. Min Act: -0.0000

Data Index: 1525328 (Python Code)

Max Activating Token Index: 1007

Text #7

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2513. Min Act: -0.0000

Data Index: 539252 (C4 (Web Text))

Max Activating Token Index: 798

Text #8

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2661. Min Act: -0.0000

Data Index: 572480 (C4 (Web Text))

Max Activating Token Index: 833

Text #9

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2704. Min Act: -0.0000

Data Index: 1406145 (Python Code)

Max Activating Token Index: 759

Text #10

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2600. Min Act: -0.0000

Data Index: 1479297 (Python Code)

Max Activating Token Index: 632

Text #11

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2567. Min Act: -0.0000

Data Index: 1588910 (Python Code)

Max Activating Token Index: 300

Text #12

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2370. Min Act: -0.0000

Data Index: 1391347 (Python Code)

Max Activating Token Index: 282

Text #13

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2489. Min Act: -0.0000

Data Index: 884667 (C4 (Web Text))

Max Activating Token Index: 333

Text #14

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2314. Min Act: -0.0000

Data Index: 80500 (C4 (Web Text))

Max Activating Token Index: 232

Text #15

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2279. Min Act: -0.0000

Data Index: 1395632 (Python Code)

Max Activating Token Index: 272

Text #16

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2317. Min Act: -0.0000

Data Index: 1370915 (Python Code)

Max Activating Token Index: 499

Text #17

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2088. Min Act: -0.0000

Data Index: 952471 (C4 (Web Text))

Max Activating Token Index: 926

Text #18

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2047. Min Act: -0.0000

Data Index: 1469452 (Python Code)

Max Activating Token Index: 295

Text #19

Max Range: 0.3099. Min Range: -0.3099

Max Act: 0.2256. Min Act: -0.0000

Data Index: 706317 (C4 (Web Text))

Max Activating Token Index: 633