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

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

Neuron 3310 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: 2.0530. Min Range: -2.0530

Max Act: 1.9767. Min Act: -0.0001

Data Index: 149753 (C4 (Web Text))

Max Activating Token Index: 96

Text #1

Max Range: 2.0530. Min Range: -2.0530

Max Act: 2.0530. Min Act: -0.0001

Data Index: 190045 (C4 (Web Text))

Max Activating Token Index: 623

Text #2

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.8927. Min Act: -0.0001

Data Index: 1077615 (C4 (Web Text))

Max Activating Token Index: 395

Text #3

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.9728. Min Act: -0.0001

Data Index: 530241 (C4 (Web Text))

Max Activating Token Index: 463

Text #4

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.8495. Min Act: -0.0001

Data Index: 1619544 (Python Code)

Max Activating Token Index: 490

Text #5

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.7332. Min Act: -0.0001

Data Index: 870829 (C4 (Web Text))

Max Activating Token Index: 770

Text #6

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.7193. Min Act: -0.0001

Data Index: 1481149 (Python Code)

Max Activating Token Index: 778

Text #7

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.6005. Min Act: -0.0000

Data Index: 1547720 (Python Code)

Max Activating Token Index: 881

Text #8

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4912. Min Act: -0.0001

Data Index: 1416264 (Python Code)

Max Activating Token Index: 917

Text #9

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.5595. Min Act: -0.0001

Data Index: 144260 (C4 (Web Text))

Max Activating Token Index: 321

Text #10

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4052. Min Act: -0.0001

Data Index: 313601 (C4 (Web Text))

Max Activating Token Index: 559

Text #11

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.5152. Min Act: -0.0001

Data Index: 1383955 (Python Code)

Max Activating Token Index: 563

Text #12

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.5038. Min Act: -0.0000

Data Index: 1430365 (Python Code)

Max Activating Token Index: 962

Text #13

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4882. Min Act: -0.0001

Data Index: 1378940 (Python Code)

Max Activating Token Index: 528

Text #14

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4608. Min Act: -0.0000

Data Index: 1457012 (Python Code)

Max Activating Token Index: 859

Text #15

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.3118. Min Act: -0.0001

Data Index: 1646969 (Python Code)

Max Activating Token Index: 708

Text #16

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4597. Min Act: -0.0000

Data Index: 1465325 (Python Code)

Max Activating Token Index: 560

Text #17

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.5131. Min Act: -0.0001

Data Index: 178125 (C4 (Web Text))

Max Activating Token Index: 139

Text #18

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.4723. Min Act: -0.0001

Data Index: 1506511 (Python Code)

Max Activating Token Index: 631

Text #19

Max Range: 2.0530. Min Range: -2.0530

Max Act: 1.3648. Min Act: -0.0001

Data Index: 1418795 (Python Code)

Max Activating Token Index: 128