Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 2552 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6460. Min Act: -0.0001

Data Index: 765002 (C4 (Web Text))

Max Activating Token Index: 176

Text #1

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6742. Min Act: -0.0001

Data Index: 625016 (C4 (Web Text))

Max Activating Token Index: 122

Text #2

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6850. Min Act: -0.0001

Data Index: 793590 (C4 (Web Text))

Max Activating Token Index: 746

Text #3

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6383. Min Act: -0.0001

Data Index: 428382 (C4 (Web Text))

Max Activating Token Index: 181

Text #4

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6435. Min Act: -0.0001

Data Index: 290375 (C4 (Web Text))

Max Activating Token Index: 27

Text #5

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6404. Min Act: -0.0001

Data Index: 718632 (C4 (Web Text))

Max Activating Token Index: 182

Text #6

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5018. Min Act: -0.0001

Data Index: 1081476 (C4 (Web Text))

Max Activating Token Index: 864

Text #7

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6059. Min Act: -0.0001

Data Index: 1132870 (C4 (Web Text))

Max Activating Token Index: 248

Text #8

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6327. Min Act: -0.0001

Data Index: 108413 (C4 (Web Text))

Max Activating Token Index: 162

Text #9

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5930. Min Act: -0.0001

Data Index: 159965 (C4 (Web Text))

Max Activating Token Index: 676

Text #10

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6314. Min Act: -0.0001

Data Index: 656697 (C4 (Web Text))

Max Activating Token Index: 742

Text #11

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6029. Min Act: -0.0001

Data Index: 169442 (C4 (Web Text))

Max Activating Token Index: 563

Text #12

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6244. Min Act: -0.0001

Data Index: 111066 (C4 (Web Text))

Max Activating Token Index: 15

Text #13

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6215. Min Act: -0.0001

Data Index: 352543 (C4 (Web Text))

Max Activating Token Index: 88

Text #14

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6073. Min Act: -0.0001

Data Index: 597135 (C4 (Web Text))

Max Activating Token Index: 621

Text #15

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.6046. Min Act: -0.0001

Data Index: 1009385 (C4 (Web Text))

Max Activating Token Index: 889

Text #16

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5982. Min Act: -0.0001

Data Index: 394949 (C4 (Web Text))

Max Activating Token Index: 829

Text #17

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5918. Min Act: -0.0001

Data Index: 1213996 (C4 (Web Text))

Max Activating Token Index: 736

Text #18

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5171. Min Act: -0.0001

Data Index: 486701 (C4 (Web Text))

Max Activating Token Index: 652

Text #19

Max Range: 0.6850. Min Range: -0.6850

Max Act: 0.5679. Min Act: -0.0001

Data Index: 1330979 (C4 (Web Text))

Max Activating Token Index: 539