Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 3805 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 4.8195. Min Range: -4.8195

Max Act: 4.8195. Min Act: -0.0000

Data Index: 708610 (C4 (Web Text))

Max Activating Token Index: 414

Text #1

Max Range: 4.8195. Min Range: -4.8195

Max Act: 4.0785. Min Act: -0.0000

Data Index: 864523 (C4 (Web Text))

Max Activating Token Index: 945

Text #2

Max Range: 4.8195. Min Range: -4.8195

Max Act: 4.0088. Min Act: -0.0000

Data Index: 987752 (C4 (Web Text))

Max Activating Token Index: 873

Text #3

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.8235. Min Act: -0.0000

Data Index: 1614246 (Python Code)

Max Activating Token Index: 112

Text #4

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.5942. Min Act: -0.0000

Data Index: 1004794 (C4 (Web Text))

Max Activating Token Index: 807

Text #5

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.9145. Min Act: -0.0000

Data Index: 1460464 (Python Code)

Max Activating Token Index: 284

Text #6

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.5926. Min Act: -0.0000

Data Index: 1399269 (Python Code)

Max Activating Token Index: 372

Text #7

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.5610. Min Act: -0.0000

Data Index: 1644546 (Python Code)

Max Activating Token Index: 601

Text #8

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.3435. Min Act: -0.0000

Data Index: 839936 (C4 (Web Text))

Max Activating Token Index: 560

Text #9

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.4814. Min Act: -0.0000

Data Index: 379547 (C4 (Web Text))

Max Activating Token Index: 1018

Text #10

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.2716. Min Act: -0.0000

Data Index: 312638 (C4 (Web Text))

Max Activating Token Index: 980

Text #11

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.0689. Min Act: -0.0000

Data Index: 1518905 (Python Code)

Max Activating Token Index: 882

Text #12

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.1503. Min Act: -0.0000

Data Index: 1382960 (Python Code)

Max Activating Token Index: 644

Text #13

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.0768. Min Act: -0.0000

Data Index: 1409930 (Python Code)

Max Activating Token Index: 714

Text #14

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.0965. Min Act: -0.0000

Data Index: 1481004 (Python Code)

Max Activating Token Index: 1011

Text #15

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.0450. Min Act: -0.0000

Data Index: 927512 (C4 (Web Text))

Max Activating Token Index: 854

Text #16

Max Range: 4.8195. Min Range: -4.8195

Max Act: 2.8790. Min Act: -0.0000

Data Index: 1392857 (Python Code)

Max Activating Token Index: 17

Text #17

Max Range: 4.8195. Min Range: -4.8195

Max Act: 3.1583. Min Act: -0.0000

Data Index: 1251354 (C4 (Web Text))

Max Activating Token Index: 698

Text #18

Max Range: 4.8195. Min Range: -4.8195

Max Act: 2.7713. Min Act: -0.0000

Data Index: 601061 (C4 (Web Text))

Max Activating Token Index: 736

Text #19

Max Range: 4.8195. Min Range: -4.8195

Max Act: 2.8670. Min Act: -0.0000

Data Index: 1419645 (Python Code)

Max Activating Token Index: 875