Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 2921 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.4620. Min Act: -0.0001

Data Index: 1211424 (C4 (Web Text))

Max Activating Token Index: 98

Text #1

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3841. Min Act: -0.0001

Data Index: 1518854 (Python Code)

Max Activating Token Index: 850

Text #2

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3559. Min Act: -0.0001

Data Index: 441754 (C4 (Web Text))

Max Activating Token Index: 180

Text #3

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3434. Min Act: -0.0001

Data Index: 382229 (C4 (Web Text))

Max Activating Token Index: 410

Text #4

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3158. Min Act: -0.0001

Data Index: 1501527 (Python Code)

Max Activating Token Index: 250

Text #5

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2913. Min Act: -0.0001

Data Index: 1616084 (Python Code)

Max Activating Token Index: 465

Text #6

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3219. Min Act: -0.0001

Data Index: 1606325 (Python Code)

Max Activating Token Index: 108

Text #7

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2955. Min Act: -0.0001

Data Index: 502634 (C4 (Web Text))

Max Activating Token Index: 543

Text #8

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3139. Min Act: -0.0001

Data Index: 1083012 (C4 (Web Text))

Max Activating Token Index: 467

Text #9

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3129. Min Act: -0.0001

Data Index: 439813 (C4 (Web Text))

Max Activating Token Index: 220

Text #10

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2999. Min Act: -0.0001

Data Index: 1381501 (Python Code)

Max Activating Token Index: 327

Text #11

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2918. Min Act: -0.0001

Data Index: 1636379 (Python Code)

Max Activating Token Index: 360

Text #12

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2938. Min Act: -0.0001

Data Index: 1679 (C4 (Web Text))

Max Activating Token Index: 1011

Text #13

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2975. Min Act: -0.0001

Data Index: 89956 (C4 (Web Text))

Max Activating Token Index: 123

Text #14

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.3071. Min Act: -0.0001

Data Index: 1457143 (Python Code)

Max Activating Token Index: 821

Text #15

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2768. Min Act: -0.0001

Data Index: 516408 (C4 (Web Text))

Max Activating Token Index: 528

Text #16

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2849. Min Act: -0.0001

Data Index: 1316096 (C4 (Web Text))

Max Activating Token Index: 411

Text #17

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2972. Min Act: -0.0001

Data Index: 2908 (C4 (Web Text))

Max Activating Token Index: 567

Text #18

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2739. Min Act: -0.0001

Data Index: 1405526 (Python Code)

Max Activating Token Index: 922

Text #19

Max Range: 0.4620. Min Range: -0.4620

Max Act: 0.2691. Min Act: -0.0001

Data Index: 381222 (C4 (Web Text))

Max Activating Token Index: 4