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

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

Neuron 638 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.4730. Min Act: -0.0001

Data Index: 1087824 (C4 (Web Text))

Max Activating Token Index: 141

Text #1

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.4979. Min Act: -0.0001

Data Index: 450089 (C4 (Web Text))

Max Activating Token Index: 114

Text #2

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.4271. Min Act: -0.0001

Data Index: 880908 (C4 (Web Text))

Max Activating Token Index: 730

Text #3

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3961. Min Act: -0.0001

Data Index: 133466 (C4 (Web Text))

Max Activating Token Index: 25

Text #4

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.4846. Min Act: -0.0001

Data Index: 1302665 (C4 (Web Text))

Max Activating Token Index: 300

Text #5

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3620. Min Act: -0.0001

Data Index: 547405 (C4 (Web Text))

Max Activating Token Index: 23

Text #6

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3657. Min Act: -0.0001

Data Index: 328299 (C4 (Web Text))

Max Activating Token Index: 139

Text #7

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3406. Min Act: -0.0001

Data Index: 891811 (C4 (Web Text))

Max Activating Token Index: 920

Text #8

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3190. Min Act: -0.0001

Data Index: 1593631 (Python Code)

Max Activating Token Index: 827

Text #9

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3326. Min Act: -0.0001

Data Index: 948222 (C4 (Web Text))

Max Activating Token Index: 162

Text #10

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3195. Min Act: -0.0001

Data Index: 771108 (C4 (Web Text))

Max Activating Token Index: 699

Text #11

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3080. Min Act: -0.0001

Data Index: 206105 (C4 (Web Text))

Max Activating Token Index: 691

Text #12

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3116. Min Act: -0.0001

Data Index: 862644 (C4 (Web Text))

Max Activating Token Index: 86

Text #13

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.2739. Min Act: -0.0001

Data Index: 1035707 (C4 (Web Text))

Max Activating Token Index: 690

Text #14

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3090. Min Act: -0.0001

Data Index: 1465683 (Python Code)

Max Activating Token Index: 501

Text #15

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3010. Min Act: -0.0001

Data Index: 1099755 (C4 (Web Text))

Max Activating Token Index: 293

Text #16

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.3106. Min Act: -0.0001

Data Index: 1041568 (C4 (Web Text))

Max Activating Token Index: 527

Text #17

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.2690. Min Act: -0.0001

Data Index: 1534490 (Python Code)

Max Activating Token Index: 440

Text #18

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.2675. Min Act: -0.0001

Data Index: 1459704 (Python Code)

Max Activating Token Index: 853

Text #19

Max Range: 0.4979. Min Range: -0.4979

Max Act: 0.2686. Min Act: -0.0001

Data Index: 1455358 (Python Code)

Max Activating Token Index: 66