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

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

Neuron 2578 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.2048. Min Act: -0.0001

Data Index: 715057 (C4 (Web Text))

Max Activating Token Index: 399

Text #1

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1976. Min Act: 0.0000

Data Index: 576103 (C4 (Web Text))

Max Activating Token Index: 415

Text #2

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.2018. Min Act: -0.0000

Data Index: 428815 (C4 (Web Text))

Max Activating Token Index: 242

Text #3

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.2018. Min Act: -0.0001

Data Index: 514222 (C4 (Web Text))

Max Activating Token Index: 729

Text #4

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1987. Min Act: 0.0000

Data Index: 65916 (C4 (Web Text))

Max Activating Token Index: 252

Text #5

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1970. Min Act: -0.0001

Data Index: 1343859 (C4 (Web Text))

Max Activating Token Index: 93

Text #6

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1870. Min Act: -0.0001

Data Index: 497186 (C4 (Web Text))

Max Activating Token Index: 316

Text #7

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1845. Min Act: 0.0000

Data Index: 648084 (C4 (Web Text))

Max Activating Token Index: 674

Text #8

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1979. Min Act: -0.0001

Data Index: 563027 (C4 (Web Text))

Max Activating Token Index: 368

Text #9

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1969. Min Act: -0.0000

Data Index: 1346503 (C4 (Web Text))

Max Activating Token Index: 268

Text #10

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1898. Min Act: 0.0000

Data Index: 628078 (C4 (Web Text))

Max Activating Token Index: 570

Text #11

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1885. Min Act: -0.0001

Data Index: 1015720 (C4 (Web Text))

Max Activating Token Index: 323

Text #12

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1841. Min Act: 0.0000

Data Index: 1035660 (C4 (Web Text))

Max Activating Token Index: 420

Text #13

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1787. Min Act: -0.0001

Data Index: 60400 (C4 (Web Text))

Max Activating Token Index: 282

Text #14

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1835. Min Act: -0.0000

Data Index: 583099 (C4 (Web Text))

Max Activating Token Index: 1019

Text #15

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1833. Min Act: -0.0000

Data Index: 1195874 (C4 (Web Text))

Max Activating Token Index: 819

Text #16

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1830. Min Act: 0.0000

Data Index: 233640 (C4 (Web Text))

Max Activating Token Index: 300

Text #17

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1851. Min Act: -0.0000

Data Index: 1271065 (C4 (Web Text))

Max Activating Token Index: 631

Text #18

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1815. Min Act: 0.0000

Data Index: 153294 (C4 (Web Text))

Max Activating Token Index: 61

Text #19

Max Range: 0.2048. Min Range: -0.2048

Max Act: 0.1752. Min Act: -0.0001

Data Index: 276631 (C4 (Web Text))

Max Activating Token Index: 889