Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 586 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0984. Min Act: -0.0003

Data Index: 832946 (C4 (Web Text))

Max Activating Token Index: 755

Text #1

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0950. Min Act: -0.0003

Data Index: 1337448 (C4 (Web Text))

Max Activating Token Index: 225

Text #2

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0974. Min Act: -0.0003

Data Index: 928331 (C4 (Web Text))

Max Activating Token Index: 202

Text #3

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0945. Min Act: -0.0003

Data Index: 1132257 (C4 (Web Text))

Max Activating Token Index: 347

Text #4

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0985. Min Act: -0.0002

Data Index: 70531 (C4 (Web Text))

Max Activating Token Index: 505

Text #5

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0925. Min Act: -0.0002

Data Index: 792494 (C4 (Web Text))

Max Activating Token Index: 674

Text #6

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0939. Min Act: -0.0003

Data Index: 1018858 (C4 (Web Text))

Max Activating Token Index: 102

Text #7

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0898. Min Act: -0.0003

Data Index: 446150 (C4 (Web Text))

Max Activating Token Index: 35

Text #8

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0925. Min Act: -0.0002

Data Index: 878186 (C4 (Web Text))

Max Activating Token Index: 672

Text #9

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0858. Min Act: -0.0003

Data Index: 1235359 (C4 (Web Text))

Max Activating Token Index: 676

Text #10

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0915. Min Act: -0.0003

Data Index: 1207623 (C4 (Web Text))

Max Activating Token Index: 964

Text #11

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0882. Min Act: -0.0003

Data Index: 1078586 (C4 (Web Text))

Max Activating Token Index: 863

Text #12

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0863. Min Act: -0.0002

Data Index: 153435 (C4 (Web Text))

Max Activating Token Index: 842

Text #13

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0859. Min Act: -0.0003

Data Index: 1084706 (C4 (Web Text))

Max Activating Token Index: 797

Text #14

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0851. Min Act: -0.0002

Data Index: 887216 (C4 (Web Text))

Max Activating Token Index: 123

Text #15

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0822. Min Act: -0.0002

Data Index: 203144 (C4 (Web Text))

Max Activating Token Index: 716

Text #16

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0826. Min Act: -0.0003

Data Index: 1138964 (C4 (Web Text))

Max Activating Token Index: 689

Text #17

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0844. Min Act: -0.0002

Data Index: 407827 (C4 (Web Text))

Max Activating Token Index: 527

Text #18

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0834. Min Act: -0.0002

Data Index: 1202960 (C4 (Web Text))

Max Activating Token Index: 506

Text #19

Max Range: 0.0985. Min Range: -0.0985

Max Act: 0.0835. Min Act: -0.0003

Data Index: 446716 (C4 (Web Text))

Max Activating Token Index: 163