Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 1640 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1160. Min Act: -0.0001

Data Index: 315637 (C4 (Web Text))

Max Activating Token Index: 549

Text #1

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1214. Min Act: -0.0001

Data Index: 1386737 (Python Code)

Max Activating Token Index: 1023

Text #2

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1045. Min Act: -0.0001

Data Index: 12381 (C4 (Web Text))

Max Activating Token Index: 138

Text #3

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1111. Min Act: -0.0001

Data Index: 189440 (C4 (Web Text))

Max Activating Token Index: 593

Text #4

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1039. Min Act: -0.0001

Data Index: 801545 (C4 (Web Text))

Max Activating Token Index: 557

Text #5

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1046. Min Act: -0.0001

Data Index: 1641239 (Python Code)

Max Activating Token Index: 489

Text #6

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1143. Min Act: -0.0001

Data Index: 532989 (C4 (Web Text))

Max Activating Token Index: 904

Text #7

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1007. Min Act: -0.0001

Data Index: 742638 (C4 (Web Text))

Max Activating Token Index: 850

Text #8

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.0990. Min Act: -0.0001

Data Index: 1411234 (Python Code)

Max Activating Token Index: 589

Text #9

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1120. Min Act: -0.0001

Data Index: 401690 (C4 (Web Text))

Max Activating Token Index: 676

Text #10

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1001. Min Act: -0.0001

Data Index: 1378011 (Python Code)

Max Activating Token Index: 758

Text #11

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1032. Min Act: -0.0000

Data Index: 1509458 (Python Code)

Max Activating Token Index: 140

Text #12

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1005. Min Act: -0.0001

Data Index: 1088027 (C4 (Web Text))

Max Activating Token Index: 343

Text #13

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1059. Min Act: -0.0001

Data Index: 1201415 (C4 (Web Text))

Max Activating Token Index: 949

Text #14

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.0969. Min Act: -0.0001

Data Index: 1647076 (Python Code)

Max Activating Token Index: 835

Text #15

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.0968. Min Act: -0.0001

Data Index: 1560007 (Python Code)

Max Activating Token Index: 995

Text #16

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1061. Min Act: -0.0001

Data Index: 1334206 (C4 (Web Text))

Max Activating Token Index: 655

Text #17

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.1000. Min Act: -0.0001

Data Index: 283485 (C4 (Web Text))

Max Activating Token Index: 566

Text #18

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.0964. Min Act: -0.0001

Data Index: 1380569 (Python Code)

Max Activating Token Index: 199

Text #19

Max Range: 0.1214. Min Range: -0.1214

Max Act: 0.0992. Min Act: -0.0001

Data Index: 1375406 (Python Code)

Max Activating Token Index: 993