Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 533 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3538. Min Act: -0.0001

Data Index: 1624772 (Python Code)

Max Activating Token Index: 612

Text #1

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3511. Min Act: -0.0001

Data Index: 1397473 (Python Code)

Max Activating Token Index: 650

Text #2

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3607. Min Act: -0.0001

Data Index: 1335154 (C4 (Web Text))

Max Activating Token Index: 631

Text #3

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3306. Min Act: -0.0001

Data Index: 234933 (C4 (Web Text))

Max Activating Token Index: 333

Text #4

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3549. Min Act: -0.0001

Data Index: 1391569 (Python Code)

Max Activating Token Index: 742

Text #5

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3391. Min Act: -0.0001

Data Index: 717839 (C4 (Web Text))

Max Activating Token Index: 660

Text #6

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3262. Min Act: -0.0001

Data Index: 789400 (C4 (Web Text))

Max Activating Token Index: 376

Text #7

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3135. Min Act: -0.0001

Data Index: 1478774 (Python Code)

Max Activating Token Index: 569

Text #8

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3485. Min Act: -0.0001

Data Index: 329920 (C4 (Web Text))

Max Activating Token Index: 1004

Text #9

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3117. Min Act: -0.0001

Data Index: 789222 (C4 (Web Text))

Max Activating Token Index: 625

Text #10

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3115. Min Act: -0.0001

Data Index: 1638510 (Python Code)

Max Activating Token Index: 627

Text #11

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3333. Min Act: -0.0001

Data Index: 786431 (C4 (Web Text))

Max Activating Token Index: 661

Text #12

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3174. Min Act: -0.0001

Data Index: 826538 (C4 (Web Text))

Max Activating Token Index: 751

Text #13

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3269. Min Act: -0.0001

Data Index: 703795 (C4 (Web Text))

Max Activating Token Index: 922

Text #14

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3244. Min Act: -0.0001

Data Index: 36593 (C4 (Web Text))

Max Activating Token Index: 13

Text #15

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3019. Min Act: -0.0001

Data Index: 1371264 (Python Code)

Max Activating Token Index: 228

Text #16

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3099. Min Act: -0.0001

Data Index: 1525793 (Python Code)

Max Activating Token Index: 766

Text #17

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3141. Min Act: -0.0001

Data Index: 909187 (C4 (Web Text))

Max Activating Token Index: 190

Text #18

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3004. Min Act: -0.0001

Data Index: 36513 (C4 (Web Text))

Max Activating Token Index: 323

Text #19

Max Range: 0.3607. Min Range: -0.3607

Max Act: 0.3104. Min Act: -0.0001

Data Index: 71940 (C4 (Web Text))

Max Activating Token Index: 370