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

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

Neuron 5104 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.4832. Min Range: -4.4832

Max Act: 4.4832. Min Act: -0.0000

Data Index: 1510744 (Python Code)

Max Activating Token Index: 853

Text #1

Max Range: 4.4832. Min Range: -4.4832

Max Act: 4.2410. Min Act: -0.0000

Data Index: 1650298 (Python Code)

Max Activating Token Index: 263

Text #2

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.9860. Min Act: -0.0000

Data Index: 1493105 (Python Code)

Max Activating Token Index: 1012

Text #3

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.8115. Min Act: -0.0000

Data Index: 1510460 (Python Code)

Max Activating Token Index: 486

Text #4

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.3193. Min Act: -0.0000

Data Index: 473312 (C4 (Web Text))

Max Activating Token Index: 133

Text #5

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.5851. Min Act: -0.0000

Data Index: 1399269 (Python Code)

Max Activating Token Index: 349

Text #6

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.5447. Min Act: -0.0000

Data Index: 1444482 (Python Code)

Max Activating Token Index: 77

Text #7

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.3931. Min Act: -0.0000

Data Index: 1045998 (C4 (Web Text))

Max Activating Token Index: 293

Text #8

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.1486. Min Act: -0.0000

Data Index: 1591314 (Python Code)

Max Activating Token Index: 588

Text #9

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.5820. Min Act: -0.0000

Data Index: 1098459 (C4 (Web Text))

Max Activating Token Index: 179

Text #10

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.4258. Min Act: -0.0000

Data Index: 1399037 (Python Code)

Max Activating Token Index: 45

Text #11

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.3972. Min Act: -0.0000

Data Index: 1539224 (Python Code)

Max Activating Token Index: 281

Text #12

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.1532. Min Act: -0.0000

Data Index: 1502975 (Python Code)

Max Activating Token Index: 554

Text #13

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.3128. Min Act: -0.0000

Data Index: 1200068 (C4 (Web Text))

Max Activating Token Index: 930

Text #14

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.3169. Min Act: -0.0000

Data Index: 266727 (C4 (Web Text))

Max Activating Token Index: 372

Text #15

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.4152. Min Act: -0.0000

Data Index: 1294664 (C4 (Web Text))

Max Activating Token Index: 142

Text #16

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.5370. Min Act: -0.0000

Data Index: 1582014 (Python Code)

Max Activating Token Index: 941

Text #17

Max Range: 4.4832. Min Range: -4.4832

Max Act: 2.9005. Min Act: -0.0000

Data Index: 62862 (C4 (Web Text))

Max Activating Token Index: 376

Text #18

Max Range: 4.4832. Min Range: -4.4832

Max Act: 2.9686. Min Act: -0.0000

Data Index: 1617351 (Python Code)

Max Activating Token Index: 905

Text #19

Max Range: 4.4832. Min Range: -4.4832

Max Act: 3.2387. Min Act: -0.0000

Data Index: 443724 (C4 (Web Text))

Max Activating Token Index: 531