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

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

Neuron 1757 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: 0.1423. Min Range: -0.1423

Max Act: 0.1423. Min Act: -0.0003

Data Index: 67735 (C4 (Web Text))

Max Activating Token Index: 178

Text #1

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1190. Min Act: -0.0003

Data Index: 1380550 (Python Code)

Max Activating Token Index: 105

Text #2

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1102. Min Act: -0.0003

Data Index: 1317802 (C4 (Web Text))

Max Activating Token Index: 562

Text #3

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1154. Min Act: -0.0003

Data Index: 631999 (C4 (Web Text))

Max Activating Token Index: 657

Text #4

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1099. Min Act: -0.0003

Data Index: 757845 (C4 (Web Text))

Max Activating Token Index: 710

Text #5

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1113. Min Act: -0.0003

Data Index: 431869 (C4 (Web Text))

Max Activating Token Index: 317

Text #6

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1051. Min Act: -0.0002

Data Index: 470231 (C4 (Web Text))

Max Activating Token Index: 924

Text #7

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1072. Min Act: -0.0004

Data Index: 258332 (C4 (Web Text))

Max Activating Token Index: 806

Text #8

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1025. Min Act: -0.0003

Data Index: 161402 (C4 (Web Text))

Max Activating Token Index: 982

Text #9

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1000. Min Act: -0.0003

Data Index: 1556403 (Python Code)

Max Activating Token Index: 766

Text #10

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0963. Min Act: -0.0003

Data Index: 1530048 (Python Code)

Max Activating Token Index: 615

Text #11

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1029. Min Act: -0.0003

Data Index: 1457911 (Python Code)

Max Activating Token Index: 111

Text #12

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1025. Min Act: -0.0003

Data Index: 804470 (C4 (Web Text))

Max Activating Token Index: 245

Text #13

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0996. Min Act: -0.0003

Data Index: 1555911 (Python Code)

Max Activating Token Index: 315

Text #14

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1013. Min Act: -0.0003

Data Index: 1499017 (Python Code)

Max Activating Token Index: 106

Text #15

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0997. Min Act: -0.0003

Data Index: 620792 (C4 (Web Text))

Max Activating Token Index: 475

Text #16

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.1010. Min Act: -0.0003

Data Index: 1613949 (Python Code)

Max Activating Token Index: 548

Text #17

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0999. Min Act: -0.0001

Data Index: 1508621 (Python Code)

Max Activating Token Index: 583

Text #18

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0949. Min Act: -0.0003

Data Index: 1073132 (C4 (Web Text))

Max Activating Token Index: 388

Text #19

Max Range: 0.1423. Min Range: -0.1423

Max Act: 0.0973. Min Act: -0.0003

Data Index: 1313382 (C4 (Web Text))

Max Activating Token Index: 283