Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: The Pile

Neuron 3471 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l-pile')

Text #0

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.4564. Min Act: -0.0000

Data Index: 268050 (The Pile)

Max Activating Token Index: 889

Text #1

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.4984. Min Act: -0.0000

Data Index: 212624 (The Pile)

Max Activating Token Index: 905

Text #2

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.1678. Min Act: -0.0000

Data Index: 449334 (The Pile)

Max Activating Token Index: 585

Text #3

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.4144. Min Act: -0.0000

Data Index: 998129 (The Pile)

Max Activating Token Index: 745

Text #4

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.0980. Min Act: -0.0000

Data Index: 1059226 (The Pile)

Max Activating Token Index: 394

Text #5

Max Range: 5.4984. Min Range: -5.4984

Max Act: 5.2290. Min Act: -0.0000

Data Index: 1533322 (The Pile)

Max Activating Token Index: 279

Text #6

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.6421. Min Act: -0.0000

Data Index: 873216 (The Pile)

Max Activating Token Index: 872

Text #7

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.7734. Min Act: -0.0000

Data Index: 1431325 (The Pile)

Max Activating Token Index: 806

Text #8

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.7708. Min Act: -0.0000

Data Index: 172264 (The Pile)

Max Activating Token Index: 668

Text #9

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.9011. Min Act: -0.0000

Data Index: 926531 (The Pile)

Max Activating Token Index: 658

Text #10

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.7562. Min Act: -0.0000

Data Index: 275010 (The Pile)

Max Activating Token Index: 679

Text #11

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.9072. Min Act: -0.0000

Data Index: 23602 (The Pile)

Max Activating Token Index: 337

Text #12

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.5877. Min Act: -0.0000

Data Index: 835669 (The Pile)

Max Activating Token Index: 534

Text #13

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.4109. Min Act: -0.0001

Data Index: 1240175 (The Pile)

Max Activating Token Index: 537

Text #14

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.5575. Min Act: -0.0000

Data Index: 1155217 (The Pile)

Max Activating Token Index: 998

Text #15

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.8066. Min Act: -0.0000

Data Index: 221070 (The Pile)

Max Activating Token Index: 799

Text #16

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.3092. Min Act: -0.0000

Data Index: 438861 (The Pile)

Max Activating Token Index: 313

Text #17

Max Range: 5.4984. Min Range: -5.4984

Max Act: 4.6540. Min Act: -0.0000

Data Index: 873953 (The Pile)

Max Activating Token Index: 577

Text #18

Max Range: 5.4984. Min Range: -5.4984

Max Act: 3.9870. Min Act: -0.0000

Data Index: 1724575 (The Pile)

Max Activating Token Index: 573

Text #19

Max Range: 5.4984. Min Range: -5.4984

Max Act: 3.8012. Min Act: -0.0000

Data Index: 1600739 (The Pile)

Max Activating Token Index: 256