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

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

Neuron 4 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: 0.1086. Min Range: -0.1086

Max Act: 0.1077. Min Act: -0.0004

Data Index: 153524 (C4 (Web Text))

Max Activating Token Index: 745

Text #1

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1086. Min Act: -0.0004

Data Index: 443788 (C4 (Web Text))

Max Activating Token Index: 875

Text #2

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1053. Min Act: -0.0003

Data Index: 1078229 (C4 (Web Text))

Max Activating Token Index: 578

Text #3

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1046. Min Act: -0.0003

Data Index: 727723 (C4 (Web Text))

Max Activating Token Index: 34

Text #4

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1034. Min Act: -0.0003

Data Index: 320042 (C4 (Web Text))

Max Activating Token Index: 630

Text #5

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1048. Min Act: -0.0003

Data Index: 1026882 (C4 (Web Text))

Max Activating Token Index: 765

Text #6

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.1002. Min Act: -0.0003

Data Index: 306765 (C4 (Web Text))

Max Activating Token Index: 721

Text #7

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0952. Min Act: -0.0003

Data Index: 350287 (C4 (Web Text))

Max Activating Token Index: 391

Text #8

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0967. Min Act: -0.0004

Data Index: 920644 (C4 (Web Text))

Max Activating Token Index: 594

Text #9

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0895. Min Act: -0.0003

Data Index: 1234330 (C4 (Web Text))

Max Activating Token Index: 836

Text #10

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0938. Min Act: -0.0003

Data Index: 1077762 (C4 (Web Text))

Max Activating Token Index: 376

Text #11

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0931. Min Act: -0.0004

Data Index: 480490 (C4 (Web Text))

Max Activating Token Index: 440

Text #12

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0902. Min Act: -0.0003

Data Index: 971956 (C4 (Web Text))

Max Activating Token Index: 859

Text #13

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0925. Min Act: -0.0003

Data Index: 1149158 (C4 (Web Text))

Max Activating Token Index: 614

Text #14

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0968. Min Act: -0.0003

Data Index: 327073 (C4 (Web Text))

Max Activating Token Index: 650

Text #15

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0892. Min Act: -0.0003

Data Index: 1094105 (C4 (Web Text))

Max Activating Token Index: 600

Text #16

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0921. Min Act: -0.0003

Data Index: 999176 (C4 (Web Text))

Max Activating Token Index: 535

Text #17

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0914. Min Act: -0.0004

Data Index: 89536 (C4 (Web Text))

Max Activating Token Index: 552

Text #18

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0902. Min Act: -0.0004

Data Index: 345543 (C4 (Web Text))

Max Activating Token Index: 127

Text #19

Max Range: 0.1086. Min Range: -0.1086

Max Act: 0.0876. Min Act: -0.0003

Data Index: 899782 (C4 (Web Text))

Max Activating Token Index: 940