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

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

Neuron 4330 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: 0.5980. Min Range: -0.5980

Max Act: 0.5980. Min Act: -0.0001

Data Index: 1031489 (C4 (Web Text))

Max Activating Token Index: 704

Text #1

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4948. Min Act: -0.0001

Data Index: 1013504 (C4 (Web Text))

Max Activating Token Index: 1005

Text #2

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4465. Min Act: -0.0001

Data Index: 181164 (C4 (Web Text))

Max Activating Token Index: 703

Text #3

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4268. Min Act: -0.0000

Data Index: 1187912 (C4 (Web Text))

Max Activating Token Index: 80

Text #4

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4193. Min Act: -0.0001

Data Index: 1536583 (Python Code)

Max Activating Token Index: 138

Text #5

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4614. Min Act: -0.0001

Data Index: 1039516 (C4 (Web Text))

Max Activating Token Index: 491

Text #6

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4303. Min Act: -0.0000

Data Index: 1027920 (C4 (Web Text))

Max Activating Token Index: 269

Text #7

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.3931. Min Act: -0.0001

Data Index: 1485160 (Python Code)

Max Activating Token Index: 69

Text #8

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4160. Min Act: -0.0001

Data Index: 1475386 (Python Code)

Max Activating Token Index: 99

Text #9

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4221. Min Act: -0.0001

Data Index: 453411 (C4 (Web Text))

Max Activating Token Index: 609

Text #10

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4270. Min Act: -0.0001

Data Index: 809556 (C4 (Web Text))

Max Activating Token Index: 363

Text #11

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.3934. Min Act: -0.0001

Data Index: 1035906 (C4 (Web Text))

Max Activating Token Index: 91

Text #12

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4124. Min Act: -0.0001

Data Index: 1529305 (Python Code)

Max Activating Token Index: 619

Text #13

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4194. Min Act: -0.0001

Data Index: 200162 (C4 (Web Text))

Max Activating Token Index: 258

Text #14

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4057. Min Act: -0.0001

Data Index: 977827 (C4 (Web Text))

Max Activating Token Index: 710

Text #15

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.3993. Min Act: -0.0001

Data Index: 1353452 (C4 (Web Text))

Max Activating Token Index: 173

Text #16

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.3899. Min Act: -0.0001

Data Index: 1514584 (Python Code)

Max Activating Token Index: 198

Text #17

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4156. Min Act: -0.0001

Data Index: 45529 (C4 (Web Text))

Max Activating Token Index: 373

Text #18

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.3866. Min Act: -0.0001

Data Index: 686436 (C4 (Web Text))

Max Activating Token Index: 577

Text #19

Max Range: 0.5980. Min Range: -0.5980

Max Act: 0.4040. Min Act: -0.0001

Data Index: 1068302 (C4 (Web Text))

Max Activating Token Index: 378