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

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

Neuron 421 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.5464. Min Act: -0.0001

Data Index: 1487270 (Python Code)

Max Activating Token Index: 636

Text #1

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.3982. Min Act: -0.0001

Data Index: 1194007 (C4 (Web Text))

Max Activating Token Index: 482

Text #2

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.4189. Min Act: -0.0001

Data Index: 1599643 (Python Code)

Max Activating Token Index: 329

Text #3

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.1358. Min Act: -0.0001

Data Index: 361938 (C4 (Web Text))

Max Activating Token Index: 291

Text #4

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.1281. Min Act: -0.0001

Data Index: 489565 (C4 (Web Text))

Max Activating Token Index: 481

Text #5

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.1306. Min Act: -0.0001

Data Index: 1432191 (Python Code)

Max Activating Token Index: 1008

Text #6

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0913. Min Act: -0.0001

Data Index: 82820 (C4 (Web Text))

Max Activating Token Index: 27

Text #7

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0553. Min Act: -0.0001

Data Index: 1322085 (C4 (Web Text))

Max Activating Token Index: 916

Text #8

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.1010. Min Act: -0.0001

Data Index: 993883 (C4 (Web Text))

Max Activating Token Index: 967

Text #9

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0647. Min Act: -0.0001

Data Index: 757178 (C4 (Web Text))

Max Activating Token Index: 887

Text #10

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0585. Min Act: -0.0001

Data Index: 63327 (C4 (Web Text))

Max Activating Token Index: 51

Text #11

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0110. Min Act: -0.0001

Data Index: 359724 (C4 (Web Text))

Max Activating Token Index: 668

Text #12

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0438. Min Act: -0.0001

Data Index: 1228325 (C4 (Web Text))

Max Activating Token Index: 882

Text #13

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0057. Min Act: -0.0001

Data Index: 148173 (C4 (Web Text))

Max Activating Token Index: 86

Text #14

Max Range: 1.5464. Min Range: -1.5464

Max Act: 0.9999. Min Act: -0.0001

Data Index: 1069824 (C4 (Web Text))

Max Activating Token Index: 440

Text #15

Max Range: 1.5464. Min Range: -1.5464

Max Act: 0.9679. Min Act: -0.0000

Data Index: 355846 (C4 (Web Text))

Max Activating Token Index: 613

Text #16

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0298. Min Act: -0.0001

Data Index: 500290 (C4 (Web Text))

Max Activating Token Index: 474

Text #17

Max Range: 1.5464. Min Range: -1.5464

Max Act: 0.9985. Min Act: -0.0001

Data Index: 252828 (C4 (Web Text))

Max Activating Token Index: 47

Text #18

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0309. Min Act: -0.0001

Data Index: 1549187 (Python Code)

Max Activating Token Index: 533

Text #19

Max Range: 1.5464. Min Range: -1.5464

Max Act: 1.0605. Min Act: -0.0001

Data Index: 706186 (C4 (Web Text))

Max Activating Token Index: 899