Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 2017 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.7321. Min Act: -0.0006

Data Index: 1569812 (Python Code)

Max Activating Token Index: 971

Text #1

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.7130. Min Act: -0.0006

Data Index: 556200 (C4 (Web Text))

Max Activating Token Index: 969

Text #2

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.6595. Min Act: -0.0006

Data Index: 1612868 (Python Code)

Max Activating Token Index: 895

Text #3

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.6007. Min Act: -0.0006

Data Index: 1212205 (C4 (Web Text))

Max Activating Token Index: 958

Text #4

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.4900. Min Act: -0.0006

Data Index: 1362968 (Python Code)

Max Activating Token Index: 792

Text #5

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.3342. Min Act: -0.0006

Data Index: 1573658 (Python Code)

Max Activating Token Index: 944

Text #6

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.2791. Min Act: -0.0006

Data Index: 1647089 (Python Code)

Max Activating Token Index: 674

Text #7

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.9487. Min Act: -0.0006

Data Index: 1560403 (Python Code)

Max Activating Token Index: 976

Text #8

Max Range: 8.7321. Min Range: -8.7321

Max Act: 8.0669. Min Act: -0.0006

Data Index: 1589275 (Python Code)

Max Activating Token Index: 784

Text #9

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.9457. Min Act: -0.0006

Data Index: 919090 (C4 (Web Text))

Max Activating Token Index: 952

Text #10

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.3884. Min Act: -0.0006

Data Index: 1591032 (Python Code)

Max Activating Token Index: 999

Text #11

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.6170. Min Act: -0.0006

Data Index: 482424 (C4 (Web Text))

Max Activating Token Index: 973

Text #12

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.3577. Min Act: -0.0006

Data Index: 1076712 (C4 (Web Text))

Max Activating Token Index: 760

Text #13

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.0797. Min Act: -0.0006

Data Index: 1421588 (Python Code)

Max Activating Token Index: 818

Text #14

Max Range: 8.7321. Min Range: -8.7321

Max Act: 7.1511. Min Act: -0.0006

Data Index: 1471251 (Python Code)

Max Activating Token Index: 920

Text #15

Max Range: 8.7321. Min Range: -8.7321

Max Act: 6.7557. Min Act: -0.0006

Data Index: 1385715 (Python Code)

Max Activating Token Index: 205

Text #16

Max Range: 8.7321. Min Range: -8.7321

Max Act: 6.9166. Min Act: -0.0006

Data Index: 1526546 (Python Code)

Max Activating Token Index: 336

Text #17

Max Range: 8.7321. Min Range: -8.7321

Max Act: 6.7444. Min Act: -0.0006

Data Index: 1520298 (Python Code)

Max Activating Token Index: 887

Text #18

Max Range: 8.7321. Min Range: -8.7321

Max Act: 6.2855. Min Act: -0.0006

Data Index: 1615344 (Python Code)

Max Activating Token Index: 1014

Text #19

Max Range: 8.7321. Min Range: -8.7321

Max Act: 6.2252. Min Act: -0.0006

Data Index: 1481823 (Python Code)

Max Activating Token Index: 921