Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 1097 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0712. Min Act: -0.0003

Data Index: 772631 (C4 (Web Text))

Max Activating Token Index: 675

Text #1

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0675. Min Act: -0.0003

Data Index: 905871 (C4 (Web Text))

Max Activating Token Index: 229

Text #2

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0671. Min Act: -0.0003

Data Index: 549077 (C4 (Web Text))

Max Activating Token Index: 78

Text #3

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0620. Min Act: -0.0003

Data Index: 1171920 (C4 (Web Text))

Max Activating Token Index: 719

Text #4

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0650. Min Act: -0.0003

Data Index: 1321500 (C4 (Web Text))

Max Activating Token Index: 366

Text #5

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0621. Min Act: -0.0003

Data Index: 509471 (C4 (Web Text))

Max Activating Token Index: 35

Text #6

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0612. Min Act: -0.0003

Data Index: 235983 (C4 (Web Text))

Max Activating Token Index: 242

Text #7

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0626. Min Act: -0.0003

Data Index: 897822 (C4 (Web Text))

Max Activating Token Index: 157

Text #8

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0620. Min Act: -0.0003

Data Index: 1532165 (Python Code)

Max Activating Token Index: 70

Text #9

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0621. Min Act: -0.0003

Data Index: 733916 (C4 (Web Text))

Max Activating Token Index: 85

Text #10

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0593. Min Act: -0.0003

Data Index: 1102858 (C4 (Web Text))

Max Activating Token Index: 945

Text #11

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0606. Min Act: -0.0003

Data Index: 338422 (C4 (Web Text))

Max Activating Token Index: 112

Text #12

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0581. Min Act: -0.0003

Data Index: 683760 (C4 (Web Text))

Max Activating Token Index: 348

Text #13

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0611. Min Act: -0.0003

Data Index: 1149953 (C4 (Web Text))

Max Activating Token Index: 446

Text #14

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0604. Min Act: -0.0003

Data Index: 707818 (C4 (Web Text))

Max Activating Token Index: 762

Text #15

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0591. Min Act: -0.0003

Data Index: 1416079 (Python Code)

Max Activating Token Index: 19

Text #16

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0603. Min Act: -0.0003

Data Index: 1550585 (Python Code)

Max Activating Token Index: 304

Text #17

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0578. Min Act: -0.0003

Data Index: 486064 (C4 (Web Text))

Max Activating Token Index: 307

Text #18

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0579. Min Act: -0.0003

Data Index: 167167 (C4 (Web Text))

Max Activating Token Index: 642

Text #19

Max Range: 0.0712. Min Range: -0.0712

Max Act: 0.0597. Min Act: -0.0003

Data Index: 458116 (C4 (Web Text))

Max Activating Token Index: 161