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

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

Neuron 1684 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.5872. Min Range: -4.5872

Max Act: 4.5872. Min Act: -0.0001

Data Index: 1019424 (C4 (Web Text))

Max Activating Token Index: 833

Text #1

Max Range: 4.5872. Min Range: -4.5872

Max Act: 4.2822. Min Act: -0.0001

Data Index: 1553973 (Python Code)

Max Activating Token Index: 960

Text #2

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.9776. Min Act: -0.0001

Data Index: 1008451 (C4 (Web Text))

Max Activating Token Index: 82

Text #3

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.7753. Min Act: -0.0001

Data Index: 1481594 (Python Code)

Max Activating Token Index: 310

Text #4

Max Range: 4.5872. Min Range: -4.5872

Max Act: 4.4623. Min Act: -0.0001

Data Index: 1163418 (C4 (Web Text))

Max Activating Token Index: 144

Text #5

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.8066. Min Act: -0.0001

Data Index: 1332528 (C4 (Web Text))

Max Activating Token Index: 717

Text #6

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.6739. Min Act: -0.0001

Data Index: 779098 (C4 (Web Text))

Max Activating Token Index: 127

Text #7

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.8437. Min Act: -0.0001

Data Index: 881208 (C4 (Web Text))

Max Activating Token Index: 917

Text #8

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.3491. Min Act: -0.0001

Data Index: 364178 (C4 (Web Text))

Max Activating Token Index: 762

Text #9

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.2993. Min Act: -0.0001

Data Index: 985837 (C4 (Web Text))

Max Activating Token Index: 659

Text #10

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.5519. Min Act: -0.0001

Data Index: 1653100 (Python Code)

Max Activating Token Index: 998

Text #11

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.5210. Min Act: -0.0000

Data Index: 1135154 (C4 (Web Text))

Max Activating Token Index: 780

Text #12

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.5019. Min Act: -0.0001

Data Index: 1194767 (C4 (Web Text))

Max Activating Token Index: 158

Text #13

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.2668. Min Act: -0.0000

Data Index: 350327 (C4 (Web Text))

Max Activating Token Index: 68

Text #14

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.4540. Min Act: -0.0001

Data Index: 926914 (C4 (Web Text))

Max Activating Token Index: 15

Text #15

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.8886. Min Act: -0.0001

Data Index: 15281 (C4 (Web Text))

Max Activating Token Index: 959

Text #16

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.2770. Min Act: -0.0001

Data Index: 111607 (C4 (Web Text))

Max Activating Token Index: 524

Text #17

Max Range: 4.5872. Min Range: -4.5872

Max Act: 3.6886. Min Act: -0.0000

Data Index: 1571744 (Python Code)

Max Activating Token Index: 835

Text #18

Max Range: 4.5872. Min Range: -4.5872

Max Act: 2.9387. Min Act: -0.0001

Data Index: 1496409 (Python Code)

Max Activating Token Index: 678

Text #19

Max Range: 4.5872. Min Range: -4.5872

Max Act: 2.9191. Min Act: -0.0001

Data Index: 1465282 (Python Code)

Max Activating Token Index: 20