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

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

Neuron 1333 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.8299. Min Act: -0.0002

Data Index: 762678 (C4 (Web Text))

Max Activating Token Index: 625

Text #1

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.8767. Min Act: -0.0002

Data Index: 1325334 (C4 (Web Text))

Max Activating Token Index: 845

Text #2

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.8233. Min Act: -0.0002

Data Index: 792338 (C4 (Web Text))

Max Activating Token Index: 147

Text #3

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.8186. Min Act: -0.0002

Data Index: 118932 (C4 (Web Text))

Max Activating Token Index: 264

Text #4

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.7727. Min Act: -0.0002

Data Index: 514835 (C4 (Web Text))

Max Activating Token Index: 529

Text #5

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6611. Min Act: -0.0002

Data Index: 743272 (C4 (Web Text))

Max Activating Token Index: 471

Text #6

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.7141. Min Act: -0.0002

Data Index: 1542686 (Python Code)

Max Activating Token Index: 912

Text #7

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6831. Min Act: -0.0002

Data Index: 1124989 (C4 (Web Text))

Max Activating Token Index: 500

Text #8

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6993. Min Act: -0.0002

Data Index: 1336728 (C4 (Web Text))

Max Activating Token Index: 560

Text #9

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.5788. Min Act: -0.0002

Data Index: 1471573 (Python Code)

Max Activating Token Index: 712

Text #10

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.7269. Min Act: -0.0002

Data Index: 122266 (C4 (Web Text))

Max Activating Token Index: 860

Text #11

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6422. Min Act: -0.0002

Data Index: 124327 (C4 (Web Text))

Max Activating Token Index: 18

Text #12

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6392. Min Act: -0.0002

Data Index: 1292578 (C4 (Web Text))

Max Activating Token Index: 23

Text #13

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.5544. Min Act: -0.0002

Data Index: 1480005 (Python Code)

Max Activating Token Index: 53

Text #14

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6421. Min Act: -0.0002

Data Index: 1591166 (Python Code)

Max Activating Token Index: 217

Text #15

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6752. Min Act: -0.0002

Data Index: 1494024 (Python Code)

Max Activating Token Index: 747

Text #16

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.5815. Min Act: -0.0002

Data Index: 240102 (C4 (Web Text))

Max Activating Token Index: 22

Text #17

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.5482. Min Act: -0.0002

Data Index: 1378839 (Python Code)

Max Activating Token Index: 1014

Text #18

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6036. Min Act: -0.0002

Data Index: 695310 (C4 (Web Text))

Max Activating Token Index: 254

Text #19

Max Range: 0.8767. Min Range: -0.8767

Max Act: 0.6608. Min Act: -0.0002

Data Index: 1529695 (Python Code)

Max Activating Token Index: 99