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

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

Neuron 3053 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.4290. Min Act: -0.0000

Data Index: 20537 (C4 (Web Text))

Max Activating Token Index: 851

Text #1

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.4764. Min Act: -0.0000

Data Index: 1494072 (Python Code)

Max Activating Token Index: 124

Text #2

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3591. Min Act: -0.0000

Data Index: 748608 (C4 (Web Text))

Max Activating Token Index: 924

Text #3

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.5242. Min Act: -0.0000

Data Index: 1126071 (C4 (Web Text))

Max Activating Token Index: 250

Text #4

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.4019. Min Act: -0.0000

Data Index: 1530556 (Python Code)

Max Activating Token Index: 345

Text #5

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3436. Min Act: -0.0000

Data Index: 1412813 (Python Code)

Max Activating Token Index: 525

Text #6

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3279. Min Act: -0.0000

Data Index: 1594225 (Python Code)

Max Activating Token Index: 766

Text #7

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3386. Min Act: -0.0000

Data Index: 1494647 (Python Code)

Max Activating Token Index: 369

Text #8

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3010. Min Act: -0.0000

Data Index: 1203978 (C4 (Web Text))

Max Activating Token Index: 802

Text #9

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3199. Min Act: -0.0000

Data Index: 574967 (C4 (Web Text))

Max Activating Token Index: 416

Text #10

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3311. Min Act: -0.0000

Data Index: 1222130 (C4 (Web Text))

Max Activating Token Index: 71

Text #11

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2988. Min Act: -0.0000

Data Index: 1632678 (Python Code)

Max Activating Token Index: 102

Text #12

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2741. Min Act: -0.0000

Data Index: 1341694 (C4 (Web Text))

Max Activating Token Index: 298

Text #13

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2987. Min Act: -0.0000

Data Index: 1488476 (Python Code)

Max Activating Token Index: 764

Text #14

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2883. Min Act: -0.0000

Data Index: 856348 (C4 (Web Text))

Max Activating Token Index: 164

Text #15

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2991. Min Act: -0.0000

Data Index: 1498446 (Python Code)

Max Activating Token Index: 84

Text #16

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2750. Min Act: -0.0000

Data Index: 347012 (C4 (Web Text))

Max Activating Token Index: 473

Text #17

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2789. Min Act: -0.0000

Data Index: 1537131 (Python Code)

Max Activating Token Index: 315

Text #18

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.2842. Min Act: -0.0000

Data Index: 790957 (C4 (Web Text))

Max Activating Token Index: 597

Text #19

Max Range: 0.5242. Min Range: -0.5242

Max Act: 0.3044. Min Act: -0.0000

Data Index: 1643421 (Python Code)

Max Activating Token Index: 970