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

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

Neuron 3654 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.7130. Min Act: -0.0001

Data Index: 1400169 (Python Code)

Max Activating Token Index: 224

Text #1

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.4974. Min Act: -0.0001

Data Index: 1513110 (Python Code)

Max Activating Token Index: 784

Text #2

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5855. Min Act: -0.0001

Data Index: 1527641 (Python Code)

Max Activating Token Index: 815

Text #3

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.6360. Min Act: -0.0001

Data Index: 1370614 (Python Code)

Max Activating Token Index: 646

Text #4

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5575. Min Act: -0.0001

Data Index: 1270215 (C4 (Web Text))

Max Activating Token Index: 209

Text #5

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5378. Min Act: -0.0001

Data Index: 1471720 (Python Code)

Max Activating Token Index: 907

Text #6

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.6301. Min Act: -0.0001

Data Index: 1335400 (C4 (Web Text))

Max Activating Token Index: 939

Text #7

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.6897. Min Act: -0.0001

Data Index: 1464083 (Python Code)

Max Activating Token Index: 620

Text #8

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5715. Min Act: -0.0002

Data Index: 769328 (C4 (Web Text))

Max Activating Token Index: 808

Text #9

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5276. Min Act: -0.0002

Data Index: 185378 (C4 (Web Text))

Max Activating Token Index: 303

Text #10

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.6439. Min Act: -0.0002

Data Index: 1098109 (C4 (Web Text))

Max Activating Token Index: 1017

Text #11

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5337. Min Act: -0.0001

Data Index: 1586820 (Python Code)

Max Activating Token Index: 969

Text #12

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5841. Min Act: -0.0001

Data Index: 1464791 (Python Code)

Max Activating Token Index: 397

Text #13

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5589. Min Act: -0.0001

Data Index: 1607303 (Python Code)

Max Activating Token Index: 169

Text #14

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5611. Min Act: -0.0002

Data Index: 229065 (C4 (Web Text))

Max Activating Token Index: 901

Text #15

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5015. Min Act: -0.0001

Data Index: 1630809 (Python Code)

Max Activating Token Index: 190

Text #16

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5598. Min Act: -0.0001

Data Index: 1526630 (Python Code)

Max Activating Token Index: 538

Text #17

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5562. Min Act: -0.0001

Data Index: 1548283 (Python Code)

Max Activating Token Index: 607

Text #18

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5484. Min Act: -0.0001

Data Index: 992181 (C4 (Web Text))

Max Activating Token Index: 48

Text #19

Max Range: 0.7130. Min Range: -0.7130

Max Act: 0.5548. Min Act: -0.0001

Data Index: 1287375 (C4 (Web Text))

Max Activating Token Index: 45