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

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

Neuron 162 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1557. Min Act: -0.0001

Data Index: 977003 (C4 (Web Text))

Max Activating Token Index: 860

Text #1

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1428. Min Act: -0.0001

Data Index: 1070551 (C4 (Web Text))

Max Activating Token Index: 78

Text #2

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1239. Min Act: -0.0001

Data Index: 1173131 (C4 (Web Text))

Max Activating Token Index: 423

Text #3

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1049. Min Act: -0.0001

Data Index: 970617 (C4 (Web Text))

Max Activating Token Index: 223

Text #4

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1076. Min Act: -0.0001

Data Index: 409812 (C4 (Web Text))

Max Activating Token Index: 698

Text #5

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1160. Min Act: -0.0001

Data Index: 10053 (C4 (Web Text))

Max Activating Token Index: 278

Text #6

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1093. Min Act: -0.0001

Data Index: 328963 (C4 (Web Text))

Max Activating Token Index: 518

Text #7

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1046. Min Act: -0.0001

Data Index: 1066787 (C4 (Web Text))

Max Activating Token Index: 115

Text #8

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1034. Min Act: -0.0001

Data Index: 130390 (C4 (Web Text))

Max Activating Token Index: 243

Text #9

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1088. Min Act: -0.0001

Data Index: 240430 (C4 (Web Text))

Max Activating Token Index: 911

Text #10

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1050. Min Act: -0.0001

Data Index: 667708 (C4 (Web Text))

Max Activating Token Index: 310

Text #11

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1055. Min Act: -0.0001

Data Index: 847060 (C4 (Web Text))

Max Activating Token Index: 198

Text #12

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0986. Min Act: -0.0001

Data Index: 640632 (C4 (Web Text))

Max Activating Token Index: 24

Text #13

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0948. Min Act: -0.0001

Data Index: 269357 (C4 (Web Text))

Max Activating Token Index: 302

Text #14

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0991. Min Act: -0.0001

Data Index: 633670 (C4 (Web Text))

Max Activating Token Index: 490

Text #15

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0976. Min Act: -0.0001

Data Index: 993557 (C4 (Web Text))

Max Activating Token Index: 714

Text #16

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0962. Min Act: -0.0001

Data Index: 1424977 (Python Code)

Max Activating Token Index: 521

Text #17

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.1015. Min Act: -0.0001

Data Index: 133451 (C4 (Web Text))

Max Activating Token Index: 274

Text #18

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0981. Min Act: -0.0001

Data Index: 823084 (C4 (Web Text))

Max Activating Token Index: 151

Text #19

Max Range: 0.1557. Min Range: -0.1557

Max Act: 0.0868. Min Act: -0.0001

Data Index: 947751 (C4 (Web Text))

Max Activating Token Index: 706