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

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

Neuron 1712 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: 0.1250. Min Range: -0.1250

Max Act: 0.1234. Min Act: -0.0001

Data Index: 415125 (C4 (Web Text))

Max Activating Token Index: 661

Text #1

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1250. Min Act: -0.0001

Data Index: 1623556 (Python Code)

Max Activating Token Index: 68

Text #2

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1247. Min Act: -0.0001

Data Index: 600471 (C4 (Web Text))

Max Activating Token Index: 864

Text #3

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1168. Min Act: -0.0001

Data Index: 196844 (C4 (Web Text))

Max Activating Token Index: 91

Text #4

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1237. Min Act: -0.0001

Data Index: 975948 (C4 (Web Text))

Max Activating Token Index: 81

Text #5

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1114. Min Act: -0.0001

Data Index: 389753 (C4 (Web Text))

Max Activating Token Index: 692

Text #6

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1181. Min Act: -0.0000

Data Index: 26684 (C4 (Web Text))

Max Activating Token Index: 913

Text #7

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1105. Min Act: -0.0001

Data Index: 70006 (C4 (Web Text))

Max Activating Token Index: 839

Text #8

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1210. Min Act: -0.0000

Data Index: 1282412 (C4 (Web Text))

Max Activating Token Index: 1008

Text #9

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1127. Min Act: -0.0000

Data Index: 77843 (C4 (Web Text))

Max Activating Token Index: 972

Text #10

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1057. Min Act: -0.0001

Data Index: 1230333 (C4 (Web Text))

Max Activating Token Index: 418

Text #11

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1016. Min Act: -0.0001

Data Index: 1572226 (Python Code)

Max Activating Token Index: 248

Text #12

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1093. Min Act: -0.0001

Data Index: 1362165 (Python Code)

Max Activating Token Index: 987

Text #13

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1031. Min Act: -0.0001

Data Index: 1000931 (C4 (Web Text))

Max Activating Token Index: 418

Text #14

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1066. Min Act: -0.0001

Data Index: 113979 (C4 (Web Text))

Max Activating Token Index: 616

Text #15

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1026. Min Act: -0.0001

Data Index: 840263 (C4 (Web Text))

Max Activating Token Index: 922

Text #16

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1054. Min Act: -0.0000

Data Index: 607177 (C4 (Web Text))

Max Activating Token Index: 931

Text #17

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1087. Min Act: -0.0001

Data Index: 937748 (C4 (Web Text))

Max Activating Token Index: 912

Text #18

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1002. Min Act: -0.0001

Data Index: 1069375 (C4 (Web Text))

Max Activating Token Index: 1014

Text #19

Max Range: 0.1250. Min Range: -0.1250

Max Act: 0.1033. Min Act: -0.0000

Data Index: 695392 (C4 (Web Text))

Max Activating Token Index: 590