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

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

Neuron 3538 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: 0.0820. Min Range: -0.0820

Max Act: 0.0813. Min Act: -0.0002

Data Index: 331292 (C4 (Web Text))

Max Activating Token Index: 779

Text #1

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0819. Min Act: -0.0002

Data Index: 652632 (C4 (Web Text))

Max Activating Token Index: 45

Text #2

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0790. Min Act: -0.0001

Data Index: 1593527 (Python Code)

Max Activating Token Index: 12

Text #3

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0787. Min Act: -0.0001

Data Index: 1613023 (Python Code)

Max Activating Token Index: 13

Text #4

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0712. Min Act: -0.0002

Data Index: 804408 (C4 (Web Text))

Max Activating Token Index: 958

Text #5

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0820. Min Act: -0.0002

Data Index: 1275194 (C4 (Web Text))

Max Activating Token Index: 653

Text #6

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0781. Min Act: -0.0001

Data Index: 1397042 (Python Code)

Max Activating Token Index: 26

Text #7

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0781. Min Act: -0.0002

Data Index: 1303901 (C4 (Web Text))

Max Activating Token Index: 1008

Text #8

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0703. Min Act: -0.0001

Data Index: 1647701 (Python Code)

Max Activating Token Index: 1015

Text #9

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0736. Min Act: -0.0001

Data Index: 1009753 (C4 (Web Text))

Max Activating Token Index: 643

Text #10

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0692. Min Act: -0.0001

Data Index: 1387721 (Python Code)

Max Activating Token Index: 707

Text #11

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0739. Min Act: -0.0001

Data Index: 1033245 (C4 (Web Text))

Max Activating Token Index: 1001

Text #12

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0764. Min Act: -0.0001

Data Index: 1456874 (Python Code)

Max Activating Token Index: 19

Text #13

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0791. Min Act: -0.0001

Data Index: 1519498 (Python Code)

Max Activating Token Index: 15

Text #14

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0788. Min Act: -0.0001

Data Index: 1640221 (Python Code)

Max Activating Token Index: 22

Text #15

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0735. Min Act: -0.0002

Data Index: 143280 (C4 (Web Text))

Max Activating Token Index: 262

Text #16

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0737. Min Act: -0.0001

Data Index: 1613496 (Python Code)

Max Activating Token Index: 30

Text #17

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0756. Min Act: -0.0002

Data Index: 1213620 (C4 (Web Text))

Max Activating Token Index: 250

Text #18

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0763. Min Act: -0.0001

Data Index: 830734 (C4 (Web Text))

Max Activating Token Index: 500

Text #19

Max Range: 0.0820. Min Range: -0.0820

Max Act: 0.0759. Min Act: -0.0001

Data Index: 327228 (C4 (Web Text))

Max Activating Token Index: 737