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

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

Neuron 3791 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: 0.8803. Min Range: -0.8803

Max Act: 0.8803. Min Act: -0.0001

Data Index: 1426628 (Python Code)

Max Activating Token Index: 620

Text #1

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.7977. Min Act: -0.0001

Data Index: 1315927 (C4 (Web Text))

Max Activating Token Index: 188

Text #2

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.7786. Min Act: -0.0001

Data Index: 1634642 (Python Code)

Max Activating Token Index: 428

Text #3

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.7298. Min Act: -0.0001

Data Index: 1468496 (Python Code)

Max Activating Token Index: 600

Text #4

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.7488. Min Act: -0.0001

Data Index: 272902 (C4 (Web Text))

Max Activating Token Index: 336

Text #5

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.7264. Min Act: -0.0001

Data Index: 1191878 (C4 (Web Text))

Max Activating Token Index: 896

Text #6

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6846. Min Act: -0.0001

Data Index: 1246777 (C4 (Web Text))

Max Activating Token Index: 1008

Text #7

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6692. Min Act: -0.0001

Data Index: 466357 (C4 (Web Text))

Max Activating Token Index: 195

Text #8

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.5981. Min Act: -0.0001

Data Index: 1606433 (Python Code)

Max Activating Token Index: 680

Text #9

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6350. Min Act: -0.0001

Data Index: 1094510 (C4 (Web Text))

Max Activating Token Index: 25

Text #10

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6316. Min Act: -0.0001

Data Index: 54423 (C4 (Web Text))

Max Activating Token Index: 879

Text #11

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6232. Min Act: -0.0001

Data Index: 671453 (C4 (Web Text))

Max Activating Token Index: 476

Text #12

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6079. Min Act: -0.0001

Data Index: 1240265 (C4 (Web Text))

Max Activating Token Index: 648

Text #13

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6296. Min Act: -0.0001

Data Index: 1615079 (Python Code)

Max Activating Token Index: 855

Text #14

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6274. Min Act: -0.0001

Data Index: 1595590 (Python Code)

Max Activating Token Index: 140

Text #15

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6070. Min Act: -0.0001

Data Index: 210277 (C4 (Web Text))

Max Activating Token Index: 708

Text #16

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6226. Min Act: -0.0001

Data Index: 1624698 (Python Code)

Max Activating Token Index: 254

Text #17

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6248. Min Act: -0.0001

Data Index: 770563 (C4 (Web Text))

Max Activating Token Index: 137

Text #18

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.6364. Min Act: -0.0001

Data Index: 539812 (C4 (Web Text))

Max Activating Token Index: 312

Text #19

Max Range: 0.8803. Min Range: -0.8803

Max Act: 0.5962. Min Act: -0.0001

Data Index: 1232159 (C4 (Web Text))

Max Activating Token Index: 471