Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: The Pile

Neuron 973 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l-pile')

Text #0

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.8837. Min Act: -0.0000

Data Index: 1577216 (The Pile)

Max Activating Token Index: 100

Text #1

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.7744. Min Act: -0.0000

Data Index: 911255 (The Pile)

Max Activating Token Index: 883

Text #2

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.3574. Min Act: -0.0000

Data Index: 528745 (The Pile)

Max Activating Token Index: 461

Text #3

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.1617. Min Act: -0.0000

Data Index: 987068 (The Pile)

Max Activating Token Index: 595

Text #4

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.1177. Min Act: -0.0000

Data Index: 447564 (The Pile)

Max Activating Token Index: 508

Text #5

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.2564. Min Act: -0.0000

Data Index: 19305 (The Pile)

Max Activating Token Index: 1009

Text #6

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.3973. Min Act: -0.0000

Data Index: 1210436 (The Pile)

Max Activating Token Index: 248

Text #7

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.0363. Min Act: -0.0000

Data Index: 90384 (The Pile)

Max Activating Token Index: 232

Text #8

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.0234. Min Act: -0.0000

Data Index: 1965169 (The Pile)

Max Activating Token Index: 684

Text #9

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.8153. Min Act: -0.0000

Data Index: 1995041 (The Pile)

Max Activating Token Index: 506

Text #10

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.1685. Min Act: -0.0000

Data Index: 25327 (The Pile)

Max Activating Token Index: 265

Text #11

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.7916. Min Act: -0.0000

Data Index: 1158055 (The Pile)

Max Activating Token Index: 20

Text #12

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.2876. Min Act: -0.0000

Data Index: 1234616 (The Pile)

Max Activating Token Index: 571

Text #13

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.7656. Min Act: -0.0000

Data Index: 683609 (The Pile)

Max Activating Token Index: 374

Text #14

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.8876. Min Act: -0.0000

Data Index: 1859273 (The Pile)

Max Activating Token Index: 40

Text #15

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.8497. Min Act: -0.0000

Data Index: 1430928 (The Pile)

Max Activating Token Index: 590

Text #16

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.8555. Min Act: -0.0000

Data Index: 1213027 (The Pile)

Max Activating Token Index: 237

Text #17

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.8582. Min Act: -0.0000

Data Index: 1527052 (The Pile)

Max Activating Token Index: 375

Text #18

Max Range: 4.8837. Min Range: -4.8837

Max Act: 4.1249. Min Act: -0.0000

Data Index: 1816734 (The Pile)

Max Activating Token Index: 169

Text #19

Max Range: 4.8837. Min Range: -4.8837

Max Act: 3.4881. Min Act: -0.0000

Data Index: 959645 (The Pile)

Max Activating Token Index: 641