Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 1814 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.6097. Min Act: -0.0001

Data Index: 879797 (C4 (Web Text))

Max Activating Token Index: 928

Text #1

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.2854. Min Act: -0.0001

Data Index: 847697 (C4 (Web Text))

Max Activating Token Index: 513

Text #2

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.2505. Min Act: -0.0001

Data Index: 1595888 (Python Code)

Max Activating Token Index: 806

Text #3

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.2284. Min Act: -0.0000

Data Index: 1289011 (C4 (Web Text))

Max Activating Token Index: 112

Text #4

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.1680. Min Act: -0.0001

Data Index: 1604514 (Python Code)

Max Activating Token Index: 292

Text #5

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.2246. Min Act: -0.0001

Data Index: 10407 (C4 (Web Text))

Max Activating Token Index: 509

Text #6

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.2288. Min Act: -0.0001

Data Index: 1224548 (C4 (Web Text))

Max Activating Token Index: 879

Text #7

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.0262. Min Act: -0.0001

Data Index: 572074 (C4 (Web Text))

Max Activating Token Index: 812

Text #8

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.8510. Min Act: -0.0001

Data Index: 1633013 (Python Code)

Max Activating Token Index: 196

Text #9

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.8284. Min Act: -0.0001

Data Index: 1424587 (Python Code)

Max Activating Token Index: 783

Text #10

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.1284. Min Act: -0.0001

Data Index: 1477850 (Python Code)

Max Activating Token Index: 519

Text #11

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.9728. Min Act: -0.0001

Data Index: 1419718 (Python Code)

Max Activating Token Index: 559

Text #12

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.8102. Min Act: -0.0000

Data Index: 1369312 (Python Code)

Max Activating Token Index: 15

Text #13

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.8209. Min Act: -0.0001

Data Index: 615768 (C4 (Web Text))

Max Activating Token Index: 615

Text #14

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.8801. Min Act: -0.0001

Data Index: 1363595 (Python Code)

Max Activating Token Index: 305

Text #15

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.7518. Min Act: -0.0001

Data Index: 1469091 (Python Code)

Max Activating Token Index: 884

Text #16

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.7361. Min Act: -0.0001

Data Index: 836193 (C4 (Web Text))

Max Activating Token Index: 574

Text #17

Max Range: 3.6097. Min Range: -3.6097

Max Act: 3.0627. Min Act: -0.0001

Data Index: 629952 (C4 (Web Text))

Max Activating Token Index: 115

Text #18

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.7940. Min Act: -0.0001

Data Index: 1353665 (C4 (Web Text))

Max Activating Token Index: 889

Text #19

Max Range: 3.6097. Min Range: -3.6097

Max Act: 2.6533. Min Act: -0.0001

Data Index: 1577007 (Python Code)

Max Activating Token Index: 758