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

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

Neuron 4904 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Click the toggle to see the full text!

Text #0

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.8729. Min Act: -0.0002

Data Index: 526709 (C4 (Web Text))

Max Activating Token Index: 45

Text #1

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.8824. Min Act: -0.0002

Data Index: 47765 (C4 (Web Text))

Max Activating Token Index: 280

Text #2

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.6065. Min Act: -0.0002

Data Index: 698914 (C4 (Web Text))

Max Activating Token Index: 794

Text #3

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5391. Min Act: -0.0002

Data Index: 475426 (C4 (Web Text))

Max Activating Token Index: 110

Text #4

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.6432. Min Act: -0.0002

Data Index: 850755 (C4 (Web Text))

Max Activating Token Index: 293

Text #5

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.6593. Min Act: -0.0002

Data Index: 90089 (C4 (Web Text))

Max Activating Token Index: 140

Text #6

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5643. Min Act: -0.0002

Data Index: 87204 (C4 (Web Text))

Max Activating Token Index: 584

Text #7

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.6217. Min Act: -0.0002

Data Index: 873158 (C4 (Web Text))

Max Activating Token Index: 282

Text #8

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5068. Min Act: -0.0002

Data Index: 767956 (C4 (Web Text))

Max Activating Token Index: 214

Text #9

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5481. Min Act: -0.0002

Data Index: 265023 (C4 (Web Text))

Max Activating Token Index: 433

Text #10

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5388. Min Act: -0.0002

Data Index: 1084062 (C4 (Web Text))

Max Activating Token Index: 35

Text #11

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.4427. Min Act: -0.0002

Data Index: 1054796 (C4 (Web Text))

Max Activating Token Index: 130

Text #12

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5271. Min Act: -0.0002

Data Index: 804762 (C4 (Web Text))

Max Activating Token Index: 3

Text #13

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5511. Min Act: -0.0002

Data Index: 601487 (C4 (Web Text))

Max Activating Token Index: 245

Text #14

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.4980. Min Act: -0.0002

Data Index: 296013 (C4 (Web Text))

Max Activating Token Index: 39

Text #15

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.4799. Min Act: -0.0002

Data Index: 1230752 (C4 (Web Text))

Max Activating Token Index: 975

Text #16

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.4829. Min Act: -0.0002

Data Index: 1525228 (Python Code)

Max Activating Token Index: 10

Text #17

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5369. Min Act: -0.0002

Data Index: 717548 (C4 (Web Text))

Max Activating Token Index: 469

Text #18

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.4456. Min Act: -0.0002

Data Index: 334620 (C4 (Web Text))

Max Activating Token Index: 3

Text #19

Max Range: 2.8824. Min Range: -2.8824

Max Act: 2.5338. Min Act: -0.0002

Data Index: 1113865 (C4 (Web Text))

Max Activating Token Index: 934