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

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

Neuron 1890 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0707. Min Act: -0.0001

Data Index: 440468 (C4 (Web Text))

Max Activating Token Index: 321

Text #1

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0700. Min Act: -0.0001

Data Index: 264416 (C4 (Web Text))

Max Activating Token Index: 556

Text #2

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0642. Min Act: -0.0001

Data Index: 647679 (C4 (Web Text))

Max Activating Token Index: 331

Text #3

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0665. Min Act: -0.0001

Data Index: 976812 (C4 (Web Text))

Max Activating Token Index: 729

Text #4

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0579. Min Act: -0.0001

Data Index: 147555 (C4 (Web Text))

Max Activating Token Index: 126

Text #5

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0575. Min Act: -0.0001

Data Index: 842618 (C4 (Web Text))

Max Activating Token Index: 462

Text #6

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0567. Min Act: -0.0001

Data Index: 274873 (C4 (Web Text))

Max Activating Token Index: 219

Text #7

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0558. Min Act: -0.0001

Data Index: 130587 (C4 (Web Text))

Max Activating Token Index: 591

Text #8

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0541. Min Act: -0.0001

Data Index: 212556 (C4 (Web Text))

Max Activating Token Index: 100

Text #9

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0559. Min Act: -0.0001

Data Index: 703792 (C4 (Web Text))

Max Activating Token Index: 488

Text #10

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0543. Min Act: -0.0001

Data Index: 1197974 (C4 (Web Text))

Max Activating Token Index: 957

Text #11

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0529. Min Act: -0.0001

Data Index: 913247 (C4 (Web Text))

Max Activating Token Index: 853

Text #12

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0507. Min Act: -0.0001

Data Index: 547621 (C4 (Web Text))

Max Activating Token Index: 705

Text #13

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0527. Min Act: -0.0001

Data Index: 93587 (C4 (Web Text))

Max Activating Token Index: 76

Text #14

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0544. Min Act: -0.0001

Data Index: 371601 (C4 (Web Text))

Max Activating Token Index: 936

Text #15

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0525. Min Act: -0.0001

Data Index: 1545599 (Python Code)

Max Activating Token Index: 368

Text #16

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0513. Min Act: -0.0001

Data Index: 384633 (C4 (Web Text))

Max Activating Token Index: 101

Text #17

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0509. Min Act: -0.0001

Data Index: 704080 (C4 (Web Text))

Max Activating Token Index: 879

Text #18

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0541. Min Act: -0.0001

Data Index: 1182333 (C4 (Web Text))

Max Activating Token Index: 936

Text #19

Max Range: 0.0707. Min Range: -0.0707

Max Act: 0.0522. Min Act: -0.0001

Data Index: 645115 (C4 (Web Text))

Max Activating Token Index: 984