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

Dataset: The Pile

Neuron 3072 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.5606. Min Range: -3.5606

Max Act: 3.5343. Min Act: -0.0000

Data Index: 751298 (The Pile)

Max Activating Token Index: 298

Text #1

Max Range: 3.5606. Min Range: -3.5606

Max Act: 3.4359. Min Act: -0.0000

Data Index: 1691340 (The Pile)

Max Activating Token Index: 42

Text #2

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.9852. Min Act: -0.0000

Data Index: 1278037 (The Pile)

Max Activating Token Index: 177

Text #3

Max Range: 3.5606. Min Range: -3.5606

Max Act: 3.5606. Min Act: -0.0000

Data Index: 1946905 (The Pile)

Max Activating Token Index: 529

Text #4

Max Range: 3.5606. Min Range: -3.5606

Max Act: 3.0330. Min Act: -0.0000

Data Index: 1757592 (The Pile)

Max Activating Token Index: 164

Text #5

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.8715. Min Act: -0.0000

Data Index: 19841 (The Pile)

Max Activating Token Index: 408

Text #6

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.6186. Min Act: -0.0000

Data Index: 953277 (The Pile)

Max Activating Token Index: 482

Text #7

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.6223. Min Act: -0.0000

Data Index: 190140 (The Pile)

Max Activating Token Index: 450

Text #8

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.7250. Min Act: -0.0000

Data Index: 1692794 (The Pile)

Max Activating Token Index: 51

Text #9

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.4981. Min Act: -0.0000

Data Index: 1844448 (The Pile)

Max Activating Token Index: 554

Text #10

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.5927. Min Act: -0.0000

Data Index: 1054755 (The Pile)

Max Activating Token Index: 751

Text #11

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.5996. Min Act: -0.0000

Data Index: 255245 (The Pile)

Max Activating Token Index: 204

Text #12

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.5740. Min Act: -0.0000

Data Index: 717537 (The Pile)

Max Activating Token Index: 371

Text #13

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.8124. Min Act: -0.0000

Data Index: 1364881 (The Pile)

Max Activating Token Index: 511

Text #14

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.5128. Min Act: -0.0000

Data Index: 1541747 (The Pile)

Max Activating Token Index: 541

Text #15

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.3767. Min Act: -0.0000

Data Index: 1850411 (The Pile)

Max Activating Token Index: 164

Text #16

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.3190. Min Act: -0.0000

Data Index: 1170969 (The Pile)

Max Activating Token Index: 289

Text #17

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.3521. Min Act: -0.0000

Data Index: 51419 (The Pile)

Max Activating Token Index: 516

Text #18

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.0977. Min Act: -0.0000

Data Index: 550912 (The Pile)

Max Activating Token Index: 554

Text #19

Max Range: 3.5606. Min Range: -3.5606

Max Act: 2.3258. Min Act: -0.0000

Data Index: 626927 (The Pile)

Max Activating Token Index: 86