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

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

Neuron 4541 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.1943. Min Act: -0.0000

Data Index: 303915 (C4 (Web Text))

Max Activating Token Index: 303

Text #1

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.3465. Min Act: -0.0000

Data Index: 299313 (C4 (Web Text))

Max Activating Token Index: 846

Text #2

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.1709. Min Act: -0.0000

Data Index: 47930 (C4 (Web Text))

Max Activating Token Index: 479

Text #3

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.0249. Min Act: -0.0000

Data Index: 1101719 (C4 (Web Text))

Max Activating Token Index: 454

Text #4

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.2301. Min Act: -0.0000

Data Index: 295815 (C4 (Web Text))

Max Activating Token Index: 284

Text #5

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9429. Min Act: -0.0000

Data Index: 1441110 (Python Code)

Max Activating Token Index: 650

Text #6

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.0428. Min Act: -0.0000

Data Index: 965450 (C4 (Web Text))

Max Activating Token Index: 886

Text #7

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9781. Min Act: -0.0000

Data Index: 1456562 (Python Code)

Max Activating Token Index: 578

Text #8

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.4365. Min Act: -0.0000

Data Index: 527176 (C4 (Web Text))

Max Activating Token Index: 753

Text #9

Max Range: 1.3465. Min Range: -1.3465

Max Act: 1.0151. Min Act: -0.0000

Data Index: 171608 (C4 (Web Text))

Max Activating Token Index: 294

Text #10

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.8957. Min Act: -0.0000

Data Index: 88986 (C4 (Web Text))

Max Activating Token Index: 321

Text #11

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9757. Min Act: -0.0000

Data Index: 1036651 (C4 (Web Text))

Max Activating Token Index: 274

Text #12

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9133. Min Act: -0.0000

Data Index: 1629215 (Python Code)

Max Activating Token Index: 110

Text #13

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.8805. Min Act: -0.0000

Data Index: 540804 (C4 (Web Text))

Max Activating Token Index: 174

Text #14

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.8996. Min Act: -0.0000

Data Index: 1221761 (C4 (Web Text))

Max Activating Token Index: 405

Text #15

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.7684. Min Act: -0.0000

Data Index: 940069 (C4 (Web Text))

Max Activating Token Index: 231

Text #16

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9465. Min Act: -0.0000

Data Index: 842984 (C4 (Web Text))

Max Activating Token Index: 186

Text #17

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.8840. Min Act: -0.0000

Data Index: 861570 (C4 (Web Text))

Max Activating Token Index: 902

Text #18

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9031. Min Act: -0.0000

Data Index: 310724 (C4 (Web Text))

Max Activating Token Index: 366

Text #19

Max Range: 1.3465. Min Range: -1.3465

Max Act: 0.9282. Min Act: -0.0000

Data Index: 765546 (C4 (Web Text))

Max Activating Token Index: 51