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

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

Neuron 4393 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.8160. Min Range: -4.8160

Max Act: 4.8160. Min Act: -0.0000

Data Index: 666871 (C4 (Web Text))

Max Activating Token Index: 25

Text #1

Max Range: 4.8160. Min Range: -4.8160

Max Act: 4.0362. Min Act: -0.0000

Data Index: 234743 (C4 (Web Text))

Max Activating Token Index: 588

Text #2

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.8825. Min Act: -0.0000

Data Index: 418454 (C4 (Web Text))

Max Activating Token Index: 851

Text #3

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.6920. Min Act: -0.0001

Data Index: 399676 (C4 (Web Text))

Max Activating Token Index: 405

Text #4

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.3309. Min Act: -0.0000

Data Index: 654282 (C4 (Web Text))

Max Activating Token Index: 853

Text #5

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.1834. Min Act: -0.0000

Data Index: 953915 (C4 (Web Text))

Max Activating Token Index: 565

Text #6

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.2158. Min Act: -0.0000

Data Index: 805131 (C4 (Web Text))

Max Activating Token Index: 871

Text #7

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.2235. Min Act: -0.0000

Data Index: 1161612 (C4 (Web Text))

Max Activating Token Index: 398

Text #8

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.1957. Min Act: -0.0000

Data Index: 745639 (C4 (Web Text))

Max Activating Token Index: 228

Text #9

Max Range: 4.8160. Min Range: -4.8160

Max Act: 3.1436. Min Act: -0.0000

Data Index: 514496 (C4 (Web Text))

Max Activating Token Index: 940

Text #10

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.7648. Min Act: -0.0000

Data Index: 832645 (C4 (Web Text))

Max Activating Token Index: 1007

Text #11

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.7617. Min Act: -0.0001

Data Index: 173367 (C4 (Web Text))

Max Activating Token Index: 228

Text #12

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.8621. Min Act: -0.0001

Data Index: 958866 (C4 (Web Text))

Max Activating Token Index: 463

Text #13

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.8007. Min Act: -0.0001

Data Index: 925735 (C4 (Web Text))

Max Activating Token Index: 22

Text #14

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.9299. Min Act: -0.0000

Data Index: 620926 (C4 (Web Text))

Max Activating Token Index: 922

Text #15

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.7713. Min Act: -0.0001

Data Index: 696551 (C4 (Web Text))

Max Activating Token Index: 878

Text #16

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.7680. Min Act: -0.0000

Data Index: 689935 (C4 (Web Text))

Max Activating Token Index: 776

Text #17

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.7032. Min Act: -0.0000

Data Index: 955638 (C4 (Web Text))

Max Activating Token Index: 441

Text #18

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.6835. Min Act: -0.0000

Data Index: 947661 (C4 (Web Text))

Max Activating Token Index: 873

Text #19

Max Range: 4.8160. Min Range: -4.8160

Max Act: 2.8027. Min Act: -0.0001

Data Index: 1134964 (C4 (Web Text))

Max Activating Token Index: 231