Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 85 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.2307. Min Act: -0.0001

Data Index: 1240874 (C4 (Web Text))

Max Activating Token Index: 452

Text #1

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.2119. Min Act: -0.0001

Data Index: 1495963 (Python Code)

Max Activating Token Index: 641

Text #2

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1743. Min Act: -0.0001

Data Index: 959836 (C4 (Web Text))

Max Activating Token Index: 274

Text #3

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1970. Min Act: -0.0001

Data Index: 56579 (C4 (Web Text))

Max Activating Token Index: 8

Text #4

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1943. Min Act: -0.0001

Data Index: 368550 (C4 (Web Text))

Max Activating Token Index: 404

Text #5

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.2021. Min Act: -0.0001

Data Index: 299899 (C4 (Web Text))

Max Activating Token Index: 411

Text #6

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1991. Min Act: -0.0001

Data Index: 565267 (C4 (Web Text))

Max Activating Token Index: 69

Text #7

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1764. Min Act: -0.0001

Data Index: 289607 (C4 (Web Text))

Max Activating Token Index: 283

Text #8

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.2110. Min Act: -0.0001

Data Index: 96512 (C4 (Web Text))

Max Activating Token Index: 67

Text #9

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1819. Min Act: -0.0001

Data Index: 981203 (C4 (Web Text))

Max Activating Token Index: 30

Text #10

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1989. Min Act: -0.0001

Data Index: 1366575 (Python Code)

Max Activating Token Index: 960

Text #11

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1806. Min Act: -0.0001

Data Index: 1173190 (C4 (Web Text))

Max Activating Token Index: 749

Text #12

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1577. Min Act: -0.0001

Data Index: 37242 (C4 (Web Text))

Max Activating Token Index: 643

Text #13

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1696. Min Act: -0.0001

Data Index: 1318829 (C4 (Web Text))

Max Activating Token Index: 272

Text #14

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1659. Min Act: -0.0001

Data Index: 1130639 (C4 (Web Text))

Max Activating Token Index: 25

Text #15

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1762. Min Act: -0.0001

Data Index: 1496787 (Python Code)

Max Activating Token Index: 393

Text #16

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1639. Min Act: -0.0001

Data Index: 135150 (C4 (Web Text))

Max Activating Token Index: 396

Text #17

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1635. Min Act: -0.0001

Data Index: 1187702 (C4 (Web Text))

Max Activating Token Index: 451

Text #18

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1671. Min Act: -0.0001

Data Index: 1456792 (Python Code)

Max Activating Token Index: 899

Text #19

Max Range: 0.2307. Min Range: -0.2307

Max Act: 0.1664. Min Act: -0.0001

Data Index: 1509881 (Python Code)

Max Activating Token Index: 479