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

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

Neuron 3652 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0330. Min Act: -0.0003

Data Index: 495904 (C4 (Web Text))

Max Activating Token Index: 1023

Text #1

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0258. Min Act: -0.0002

Data Index: 1468491 (Python Code)

Max Activating Token Index: 1023

Text #2

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0276. Min Act: -0.0002

Data Index: 868122 (C4 (Web Text))

Max Activating Token Index: 1023

Text #3

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0247. Min Act: -0.0003

Data Index: 862022 (C4 (Web Text))

Max Activating Token Index: 1023

Text #4

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0224. Min Act: -0.0003

Data Index: 572831 (C4 (Web Text))

Max Activating Token Index: 1023

Text #5

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0222. Min Act: -0.0002

Data Index: 749374 (C4 (Web Text))

Max Activating Token Index: 1023

Text #6

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0217. Min Act: -0.0003

Data Index: 1391514 (Python Code)

Max Activating Token Index: 83

Text #7

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0247. Min Act: -0.0003

Data Index: 463951 (C4 (Web Text))

Max Activating Token Index: 1023

Text #8

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0219. Min Act: -0.0002

Data Index: 632630 (C4 (Web Text))

Max Activating Token Index: 1023

Text #9

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0220. Min Act: -0.0003

Data Index: 1569730 (Python Code)

Max Activating Token Index: 95

Text #10

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0207. Min Act: -0.0003

Data Index: 108299 (C4 (Web Text))

Max Activating Token Index: 1023

Text #11

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0250. Min Act: -0.0003

Data Index: 179118 (C4 (Web Text))

Max Activating Token Index: 682

Text #12

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0201. Min Act: -0.0002

Data Index: 27897 (C4 (Web Text))

Max Activating Token Index: 942

Text #13

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0164. Min Act: -0.0002

Data Index: 850814 (C4 (Web Text))

Max Activating Token Index: 748

Text #14

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0202. Min Act: -0.0002

Data Index: 1641152 (Python Code)

Max Activating Token Index: 147

Text #15

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0177. Min Act: -0.0003

Data Index: 222839 (C4 (Web Text))

Max Activating Token Index: 132

Text #16

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0199. Min Act: -0.0002

Data Index: 918199 (C4 (Web Text))

Max Activating Token Index: 705

Text #17

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0205. Min Act: -0.0002

Data Index: 802345 (C4 (Web Text))

Max Activating Token Index: 834

Text #18

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0201. Min Act: -0.0003

Data Index: 629045 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 0.0330. Min Range: -0.0330

Max Act: 0.0198. Min Act: -0.0002

Data Index: 1476144 (Python Code)

Max Activating Token Index: 883