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

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

Neuron 3637 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3196. Min Act: -0.0000

Data Index: 343758 (C4 (Web Text))

Max Activating Token Index: 276

Text #1

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3120. Min Act: -0.0000

Data Index: 1544667 (Python Code)

Max Activating Token Index: 274

Text #2

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3146. Min Act: -0.0000

Data Index: 989257 (C4 (Web Text))

Max Activating Token Index: 167

Text #3

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2983. Min Act: -0.0000

Data Index: 649857 (C4 (Web Text))

Max Activating Token Index: 603

Text #4

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3047. Min Act: -0.0000

Data Index: 1154 (C4 (Web Text))

Max Activating Token Index: 131

Text #5

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2998. Min Act: -0.0000

Data Index: 1295111 (C4 (Web Text))

Max Activating Token Index: 267

Text #6

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3014. Min Act: -0.0000

Data Index: 1501747 (Python Code)

Max Activating Token Index: 370

Text #7

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2857. Min Act: -0.0000

Data Index: 137468 (C4 (Web Text))

Max Activating Token Index: 670

Text #8

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2770. Min Act: -0.0000

Data Index: 792006 (C4 (Web Text))

Max Activating Token Index: 845

Text #9

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.3093. Min Act: -0.0000

Data Index: 1505819 (Python Code)

Max Activating Token Index: 665

Text #10

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2972. Min Act: -0.0000

Data Index: 773428 (C4 (Web Text))

Max Activating Token Index: 853

Text #11

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2734. Min Act: -0.0000

Data Index: 133685 (C4 (Web Text))

Max Activating Token Index: 1010

Text #12

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2845. Min Act: -0.0000

Data Index: 341650 (C4 (Web Text))

Max Activating Token Index: 47

Text #13

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2736. Min Act: -0.0000

Data Index: 231295 (C4 (Web Text))

Max Activating Token Index: 476

Text #14

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2722. Min Act: -0.0000

Data Index: 83095 (C4 (Web Text))

Max Activating Token Index: 860

Text #15

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2716. Min Act: -0.0000

Data Index: 628672 (C4 (Web Text))

Max Activating Token Index: 370

Text #16

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2710. Min Act: 0.0000

Data Index: 577127 (C4 (Web Text))

Max Activating Token Index: 832

Text #17

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2853. Min Act: -0.0000

Data Index: 1426655 (Python Code)

Max Activating Token Index: 554

Text #18

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2768. Min Act: -0.0000

Data Index: 81534 (C4 (Web Text))

Max Activating Token Index: 270

Text #19

Max Range: 0.3196. Min Range: -0.3196

Max Act: 0.2790. Min Act: 0.0001

Data Index: 1642511 (Python Code)

Max Activating Token Index: 24