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

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

Neuron 4535 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.3497. Min Act: -0.0000

Data Index: 318048 (C4 (Web Text))

Max Activating Token Index: 937

Text #1

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.1434. Min Act: -0.0000

Data Index: 758038 (C4 (Web Text))

Max Activating Token Index: 224

Text #2

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.0186. Min Act: -0.0000

Data Index: 391490 (C4 (Web Text))

Max Activating Token Index: 946

Text #3

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.1494. Min Act: -0.0000

Data Index: 501010 (C4 (Web Text))

Max Activating Token Index: 184

Text #4

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.0055. Min Act: -0.0000

Data Index: 1143300 (C4 (Web Text))

Max Activating Token Index: 916

Text #5

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.1653. Min Act: -0.0000

Data Index: 600992 (C4 (Web Text))

Max Activating Token Index: 115

Text #6

Max Range: 1.3497. Min Range: -1.3497

Max Act: 1.0523. Min Act: -0.0000

Data Index: 694455 (C4 (Web Text))

Max Activating Token Index: 991

Text #7

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.9640. Min Act: -0.0000

Data Index: 1280612 (C4 (Web Text))

Max Activating Token Index: 473

Text #8

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.7360. Min Act: -0.0000

Data Index: 1407734 (Python Code)

Max Activating Token Index: 824

Text #9

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8510. Min Act: -0.0000

Data Index: 819060 (C4 (Web Text))

Max Activating Token Index: 760

Text #10

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.9465. Min Act: -0.0000

Data Index: 159608 (C4 (Web Text))

Max Activating Token Index: 301

Text #11

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.9623. Min Act: -0.0000

Data Index: 1004042 (C4 (Web Text))

Max Activating Token Index: 292

Text #12

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8949. Min Act: -0.0000

Data Index: 291360 (C4 (Web Text))

Max Activating Token Index: 820

Text #13

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8967. Min Act: -0.0000

Data Index: 1547870 (Python Code)

Max Activating Token Index: 143

Text #14

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8275. Min Act: -0.0000

Data Index: 1069951 (C4 (Web Text))

Max Activating Token Index: 944

Text #15

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8015. Min Act: -0.0000

Data Index: 1463809 (Python Code)

Max Activating Token Index: 20

Text #16

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.8994. Min Act: -0.0000

Data Index: 276876 (C4 (Web Text))

Max Activating Token Index: 956

Text #17

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.7856. Min Act: -0.0000

Data Index: 932642 (C4 (Web Text))

Max Activating Token Index: 541

Text #18

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.7696. Min Act: -0.0000

Data Index: 1647271 (Python Code)

Max Activating Token Index: 553

Text #19

Max Range: 1.3497. Min Range: -1.3497

Max Act: 0.7984. Min Act: -0.0000

Data Index: 1647914 (Python Code)

Max Activating Token Index: 302