Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2218 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.5697. Min Act: -0.0000

Data Index: 496968 (C4 (Web Text))

Max Activating Token Index: 380

Text #1

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.5616. Min Act: -0.0000

Data Index: 158062 (C4 (Web Text))

Max Activating Token Index: 391

Text #2

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.5325. Min Act: -0.0000

Data Index: 459095 (C4 (Web Text))

Max Activating Token Index: 32

Text #3

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.5127. Min Act: -0.0000

Data Index: 828730 (C4 (Web Text))

Max Activating Token Index: 35

Text #4

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.5047. Min Act: -0.0001

Data Index: 1268929 (C4 (Web Text))

Max Activating Token Index: 219

Text #5

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4837. Min Act: 0.0001

Data Index: 825099 (C4 (Web Text))

Max Activating Token Index: 235

Text #6

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4723. Min Act: -0.0000

Data Index: 50139 (C4 (Web Text))

Max Activating Token Index: 352

Text #7

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4846. Min Act: -0.0000

Data Index: 872263 (C4 (Web Text))

Max Activating Token Index: 61

Text #8

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4484. Min Act: -0.0000

Data Index: 491006 (C4 (Web Text))

Max Activating Token Index: 433

Text #9

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4662. Min Act: -0.0000

Data Index: 1152746 (C4 (Web Text))

Max Activating Token Index: 1012

Text #10

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4778. Min Act: -0.0000

Data Index: 99361 (C4 (Web Text))

Max Activating Token Index: 564

Text #11

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4350. Min Act: -0.0000

Data Index: 1059033 (C4 (Web Text))

Max Activating Token Index: 797

Text #12

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4479. Min Act: -0.0000

Data Index: 217430 (C4 (Web Text))

Max Activating Token Index: 940

Text #13

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4546. Min Act: -0.0000

Data Index: 868110 (C4 (Web Text))

Max Activating Token Index: 559

Text #14

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4216. Min Act: -0.0001

Data Index: 835754 (C4 (Web Text))

Max Activating Token Index: 834

Text #15

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4513. Min Act: -0.0000

Data Index: 857153 (C4 (Web Text))

Max Activating Token Index: 932

Text #16

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4309. Min Act: -0.0000

Data Index: 108333 (C4 (Web Text))

Max Activating Token Index: 59

Text #17

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4291. Min Act: -0.0001

Data Index: 258369 (C4 (Web Text))

Max Activating Token Index: 724

Text #18

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4126. Min Act: -0.0000

Data Index: 408718 (C4 (Web Text))

Max Activating Token Index: 15

Text #19

Max Range: 0.5697. Min Range: -0.5697

Max Act: 0.4230. Min Act: -0.0000

Data Index: 317790 (C4 (Web Text))

Max Activating Token Index: 480