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

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

Neuron 3951 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: 0.9887. Min Range: -0.9887

Max Act: 0.9887. Min Act: -0.0001

Data Index: 1418833 (Python Code)

Max Activating Token Index: 977

Text #1

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8944. Min Act: -0.0000

Data Index: 256517 (C4 (Web Text))

Max Activating Token Index: 293

Text #2

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.9071. Min Act: -0.0001

Data Index: 518606 (C4 (Web Text))

Max Activating Token Index: 469

Text #3

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8846. Min Act: -0.0001

Data Index: 73642 (C4 (Web Text))

Max Activating Token Index: 941

Text #4

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8950. Min Act: -0.0000

Data Index: 1280879 (C4 (Web Text))

Max Activating Token Index: 934

Text #5

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8621. Min Act: -0.0000

Data Index: 763989 (C4 (Web Text))

Max Activating Token Index: 389

Text #6

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8606. Min Act: -0.0001

Data Index: 15688 (C4 (Web Text))

Max Activating Token Index: 257

Text #7

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.9108. Min Act: -0.0000

Data Index: 183259 (C4 (Web Text))

Max Activating Token Index: 941

Text #8

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8472. Min Act: -0.0001

Data Index: 361112 (C4 (Web Text))

Max Activating Token Index: 520

Text #9

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8400. Min Act: -0.0001

Data Index: 265444 (C4 (Web Text))

Max Activating Token Index: 20

Text #10

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8356. Min Act: -0.0000

Data Index: 883388 (C4 (Web Text))

Max Activating Token Index: 901

Text #11

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8292. Min Act: -0.0000

Data Index: 508625 (C4 (Web Text))

Max Activating Token Index: 102

Text #12

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8256. Min Act: -0.0001

Data Index: 430427 (C4 (Web Text))

Max Activating Token Index: 849

Text #13

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8228. Min Act: -0.0000

Data Index: 1059318 (C4 (Web Text))

Max Activating Token Index: 189

Text #14

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8178. Min Act: -0.0001

Data Index: 1537576 (Python Code)

Max Activating Token Index: 33

Text #15

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8200. Min Act: -0.0000

Data Index: 1575163 (Python Code)

Max Activating Token Index: 909

Text #16

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8336. Min Act: -0.0001

Data Index: 510788 (C4 (Web Text))

Max Activating Token Index: 320

Text #17

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8102. Min Act: -0.0000

Data Index: 1324760 (C4 (Web Text))

Max Activating Token Index: 971

Text #18

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.8329. Min Act: -0.0000

Data Index: 624304 (C4 (Web Text))

Max Activating Token Index: 402

Text #19

Max Range: 0.9887. Min Range: -0.9887

Max Act: 0.7744. Min Act: -0.0000

Data Index: 144717 (C4 (Web Text))

Max Activating Token Index: 575