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

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

Neuron 1350 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.3007. Min Act: -0.0000

Data Index: 325699 (C4 (Web Text))

Max Activating Token Index: 785

Text #1

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.2538. Min Act: -0.0000

Data Index: 910786 (C4 (Web Text))

Max Activating Token Index: 761

Text #2

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.1654. Min Act: -0.0000

Data Index: 32309 (C4 (Web Text))

Max Activating Token Index: 879

Text #3

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.7656. Min Act: -0.0000

Data Index: 1236325 (C4 (Web Text))

Max Activating Token Index: 364

Text #4

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.1221. Min Act: -0.0000

Data Index: 1227949 (C4 (Web Text))

Max Activating Token Index: 478

Text #5

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.1276. Min Act: -0.0000

Data Index: 965516 (C4 (Web Text))

Max Activating Token Index: 320

Text #6

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.2913. Min Act: -0.0000

Data Index: 1069612 (C4 (Web Text))

Max Activating Token Index: 391

Text #7

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.1095. Min Act: -0.0000

Data Index: 498755 (C4 (Web Text))

Max Activating Token Index: 425

Text #8

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.9044. Min Act: -0.0000

Data Index: 42540 (C4 (Web Text))

Max Activating Token Index: 974

Text #9

Max Range: 7.3007. Min Range: -7.3007

Max Act: 7.0256. Min Act: -0.0000

Data Index: 162737 (C4 (Web Text))

Max Activating Token Index: 206

Text #10

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.5673. Min Act: -0.0000

Data Index: 484085 (C4 (Web Text))

Max Activating Token Index: 295

Text #11

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.5501. Min Act: -0.0000

Data Index: 922897 (C4 (Web Text))

Max Activating Token Index: 261

Text #12

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.8761. Min Act: -0.0000

Data Index: 900816 (C4 (Web Text))

Max Activating Token Index: 726

Text #13

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.5351. Min Act: -0.0000

Data Index: 1330367 (C4 (Web Text))

Max Activating Token Index: 111

Text #14

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.5759. Min Act: -0.0000

Data Index: 1277888 (C4 (Web Text))

Max Activating Token Index: 194

Text #15

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.4261. Min Act: -0.0000

Data Index: 964810 (C4 (Web Text))

Max Activating Token Index: 566

Text #16

Max Range: 7.3007. Min Range: -7.3007

Max Act: 6.1992. Min Act: -0.0000

Data Index: 937200 (C4 (Web Text))

Max Activating Token Index: 47

Text #17

Max Range: 7.3007. Min Range: -7.3007

Max Act: 5.2487. Min Act: -0.0000

Data Index: 1540687 (Python Code)

Max Activating Token Index: 29

Text #18

Max Range: 7.3007. Min Range: -7.3007

Max Act: 5.2237. Min Act: -0.0000

Data Index: 146847 (C4 (Web Text))

Max Activating Token Index: 970

Text #19

Max Range: 7.3007. Min Range: -7.3007

Max Act: 5.2568. Min Act: -0.0000

Data Index: 1036358 (C4 (Web Text))

Max Activating Token Index: 912