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

Dataset: The Pile

Neuron 1059 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.4233. Min Act: -0.0000

Data Index: 1067459 (The Pile)

Max Activating Token Index: 19

Text #1

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.1474. Min Act: -0.0000

Data Index: 1982759 (The Pile)

Max Activating Token Index: 989

Text #2

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.2487. Min Act: -0.0000

Data Index: 1335745 (The Pile)

Max Activating Token Index: 19

Text #3

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.0389. Min Act: -0.0000

Data Index: 1316725 (The Pile)

Max Activating Token Index: 639

Text #4

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.1082. Min Act: -0.0000

Data Index: 77574 (The Pile)

Max Activating Token Index: 108

Text #5

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.0902. Min Act: -0.0000

Data Index: 871350 (The Pile)

Max Activating Token Index: 239

Text #6

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.9237. Min Act: -0.0000

Data Index: 846011 (The Pile)

Max Activating Token Index: 822

Text #7

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.6053. Min Act: -0.0000

Data Index: 86803 (The Pile)

Max Activating Token Index: 986

Text #8

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.9126. Min Act: -0.0000

Data Index: 1029480 (The Pile)

Max Activating Token Index: 392

Text #9

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.0607. Min Act: -0.0000

Data Index: 326989 (The Pile)

Max Activating Token Index: 369

Text #10

Max Range: 9.4233. Min Range: -9.4233

Max Act: 9.0446. Min Act: -0.0000

Data Index: 194612 (The Pile)

Max Activating Token Index: 294

Text #11

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.9028. Min Act: -0.0000

Data Index: 1753383 (The Pile)

Max Activating Token Index: 303

Text #12

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.8793. Min Act: -0.0000

Data Index: 1850569 (The Pile)

Max Activating Token Index: 100

Text #13

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.8476. Min Act: -0.0000

Data Index: 52791 (The Pile)

Max Activating Token Index: 746

Text #14

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.7968. Min Act: -0.0000

Data Index: 1643369 (The Pile)

Max Activating Token Index: 456

Text #15

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.6380. Min Act: -0.0000

Data Index: 805823 (The Pile)

Max Activating Token Index: 189

Text #16

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.9528. Min Act: -0.0000

Data Index: 698386 (The Pile)

Max Activating Token Index: 228

Text #17

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.6593. Min Act: -0.0000

Data Index: 1831865 (The Pile)

Max Activating Token Index: 17

Text #18

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.7486. Min Act: -0.0000

Data Index: 1278612 (The Pile)

Max Activating Token Index: 134

Text #19

Max Range: 9.4233. Min Range: -9.4233

Max Act: 8.5682. Min Act: -0.0000

Data Index: 861490 (The Pile)

Max Activating Token Index: 18