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

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

Neuron 3808 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Click the toggle to see the full text!

Text #0

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1742. Min Act: -0.0001

Data Index: 371485 (C4 (Web Text))

Max Activating Token Index: 1023

Text #1

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1726. Min Act: -0.0001

Data Index: 91366 (C4 (Web Text))

Max Activating Token Index: 1023

Text #2

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1717. Min Act: -0.0001

Data Index: 537368 (C4 (Web Text))

Max Activating Token Index: 1023

Text #3

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1692. Min Act: -0.0001

Data Index: 1248038 (C4 (Web Text))

Max Activating Token Index: 1023

Text #4

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1662. Min Act: -0.0001

Data Index: 1430406 (Python Code)

Max Activating Token Index: 2

Text #5

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1670. Min Act: -0.0001

Data Index: 965035 (C4 (Web Text))

Max Activating Token Index: 1023

Text #6

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1662. Min Act: -0.0001

Data Index: 1184504 (C4 (Web Text))

Max Activating Token Index: 1023

Text #7

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1657. Min Act: -0.0001

Data Index: 1293848 (C4 (Web Text))

Max Activating Token Index: 1023

Text #8

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1653. Min Act: -0.0001

Data Index: 1565575 (Python Code)

Max Activating Token Index: 668

Text #9

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1634. Min Act: -0.0002

Data Index: 1607672 (Python Code)

Max Activating Token Index: 4

Text #10

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1611. Min Act: -0.0001

Data Index: 839176 (C4 (Web Text))

Max Activating Token Index: 2

Text #11

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1611. Min Act: -0.0001

Data Index: 953089 (C4 (Web Text))

Max Activating Token Index: 2

Text #12

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1655. Min Act: -0.0001

Data Index: 1503166 (Python Code)

Max Activating Token Index: 635

Text #13

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1644. Min Act: -0.0001

Data Index: 400972 (C4 (Web Text))

Max Activating Token Index: 2

Text #14

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1616. Min Act: -0.0001

Data Index: 370961 (C4 (Web Text))

Max Activating Token Index: 1023

Text #15

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1611. Min Act: -0.0001

Data Index: 1147194 (C4 (Web Text))

Max Activating Token Index: 1023

Text #16

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1601. Min Act: -0.0001

Data Index: 1595490 (Python Code)

Max Activating Token Index: 113

Text #17

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1612. Min Act: -0.0001

Data Index: 1087950 (C4 (Web Text))

Max Activating Token Index: 1023

Text #18

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1635. Min Act: -0.0001

Data Index: 1034233 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 0.1742. Min Range: -0.1742

Max Act: 0.1626. Min Act: -0.0001

Data Index: 200252 (C4 (Web Text))

Max Activating Token Index: 1023