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

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

Neuron 4612 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: 1.5778. Min Range: -1.5778

Max Act: 1.5778. Min Act: -0.0000

Data Index: 91516 (C4 (Web Text))

Max Activating Token Index: 112

Text #1

Max Range: 1.5778. Min Range: -1.5778

Max Act: 1.3653. Min Act: -0.0000

Data Index: 1443594 (Python Code)

Max Activating Token Index: 291

Text #2

Max Range: 1.5778. Min Range: -1.5778

Max Act: 1.3974. Min Act: -0.0000

Data Index: 42286 (C4 (Web Text))

Max Activating Token Index: 996

Text #3

Max Range: 1.5778. Min Range: -1.5778

Max Act: 1.0554. Min Act: -0.0000

Data Index: 1641110 (Python Code)

Max Activating Token Index: 307

Text #4

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.9604. Min Act: -0.0000

Data Index: 1449352 (Python Code)

Max Activating Token Index: 158

Text #5

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.9051. Min Act: -0.0000

Data Index: 106713 (C4 (Web Text))

Max Activating Token Index: 332

Text #6

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8489. Min Act: -0.0000

Data Index: 101556 (C4 (Web Text))

Max Activating Token Index: 521

Text #7

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8827. Min Act: -0.0000

Data Index: 877959 (C4 (Web Text))

Max Activating Token Index: 514

Text #8

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8292. Min Act: -0.0000

Data Index: 808047 (C4 (Web Text))

Max Activating Token Index: 609

Text #9

Max Range: 1.5778. Min Range: -1.5778

Max Act: 1.0420. Min Act: -0.0000

Data Index: 1498143 (Python Code)

Max Activating Token Index: 246

Text #10

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8477. Min Act: -0.0000

Data Index: 653972 (C4 (Web Text))

Max Activating Token Index: 215

Text #11

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8351. Min Act: -0.0000

Data Index: 857839 (C4 (Web Text))

Max Activating Token Index: 946

Text #12

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.7951. Min Act: -0.0000

Data Index: 404752 (C4 (Web Text))

Max Activating Token Index: 726

Text #13

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8427. Min Act: -0.0000

Data Index: 1368468 (Python Code)

Max Activating Token Index: 258

Text #14

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.7882. Min Act: -0.0000

Data Index: 93469 (C4 (Web Text))

Max Activating Token Index: 386

Text #15

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8143. Min Act: -0.0000

Data Index: 1308322 (C4 (Web Text))

Max Activating Token Index: 897

Text #16

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8337. Min Act: -0.0000

Data Index: 97814 (C4 (Web Text))

Max Activating Token Index: 187

Text #17

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.7977. Min Act: -0.0000

Data Index: 1606186 (Python Code)

Max Activating Token Index: 104

Text #18

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.6516. Min Act: -0.0000

Data Index: 96392 (C4 (Web Text))

Max Activating Token Index: 929

Text #19

Max Range: 1.5778. Min Range: -1.5778

Max Act: 0.8435. Min Act: -0.0000

Data Index: 550805 (C4 (Web Text))

Max Activating Token Index: 680