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

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

Neuron 4065 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.5710. Min Act: -0.0001

Data Index: 856950 (C4 (Web Text))

Max Activating Token Index: 337

Text #1

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.5176. Min Act: -0.0001

Data Index: 888744 (C4 (Web Text))

Max Activating Token Index: 979

Text #2

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4986. Min Act: -0.0001

Data Index: 977660 (C4 (Web Text))

Max Activating Token Index: 758

Text #3

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.5096. Min Act: -0.0001

Data Index: 943333 (C4 (Web Text))

Max Activating Token Index: 953

Text #4

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4426. Min Act: -0.0001

Data Index: 238741 (C4 (Web Text))

Max Activating Token Index: 298

Text #5

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.5372. Min Act: -0.0001

Data Index: 1267785 (C4 (Web Text))

Max Activating Token Index: 555

Text #6

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4837. Min Act: -0.0001

Data Index: 1277765 (C4 (Web Text))

Max Activating Token Index: 57

Text #7

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4653. Min Act: -0.0001

Data Index: 235101 (C4 (Web Text))

Max Activating Token Index: 908

Text #8

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4624. Min Act: -0.0001

Data Index: 1130649 (C4 (Web Text))

Max Activating Token Index: 253

Text #9

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4556. Min Act: -0.0001

Data Index: 607271 (C4 (Web Text))

Max Activating Token Index: 805

Text #10

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4852. Min Act: -0.0001

Data Index: 775328 (C4 (Web Text))

Max Activating Token Index: 455

Text #11

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.5148. Min Act: -0.0001

Data Index: 137255 (C4 (Web Text))

Max Activating Token Index: 727

Text #12

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4943. Min Act: -0.0001

Data Index: 107171 (C4 (Web Text))

Max Activating Token Index: 323

Text #13

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4465. Min Act: -0.0001

Data Index: 803182 (C4 (Web Text))

Max Activating Token Index: 310

Text #14

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4321. Min Act: -0.0001

Data Index: 344884 (C4 (Web Text))

Max Activating Token Index: 951

Text #15

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4470. Min Act: -0.0001

Data Index: 1631329 (Python Code)

Max Activating Token Index: 646

Text #16

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4362. Min Act: -0.0001

Data Index: 1157356 (C4 (Web Text))

Max Activating Token Index: 260

Text #17

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4938. Min Act: -0.0001

Data Index: 543330 (C4 (Web Text))

Max Activating Token Index: 656

Text #18

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4307. Min Act: -0.0001

Data Index: 225071 (C4 (Web Text))

Max Activating Token Index: 79

Text #19

Max Range: 0.5710. Min Range: -0.5710

Max Act: 0.4572. Min Act: -0.0001

Data Index: 1328953 (C4 (Web Text))

Max Activating Token Index: 857