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

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

Neuron 3829 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2743. Min Act: -0.0000

Data Index: 1060046 (C4 (Web Text))

Max Activating Token Index: 990

Text #1

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2620. Min Act: -0.0000

Data Index: 906101 (C4 (Web Text))

Max Activating Token Index: 574

Text #2

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2750. Min Act: -0.0000

Data Index: 285499 (C4 (Web Text))

Max Activating Token Index: 77

Text #3

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2663. Min Act: -0.0000

Data Index: 1071890 (C4 (Web Text))

Max Activating Token Index: 1008

Text #4

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2500. Min Act: -0.0000

Data Index: 714803 (C4 (Web Text))

Max Activating Token Index: 125

Text #5

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2481. Min Act: -0.0000

Data Index: 271443 (C4 (Web Text))

Max Activating Token Index: 106

Text #6

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2440. Min Act: -0.0000

Data Index: 647626 (C4 (Web Text))

Max Activating Token Index: 395

Text #7

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2424. Min Act: -0.0000

Data Index: 510123 (C4 (Web Text))

Max Activating Token Index: 95

Text #8

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2423. Min Act: -0.0000

Data Index: 508090 (C4 (Web Text))

Max Activating Token Index: 571

Text #9

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2215. Min Act: -0.0000

Data Index: 158655 (C4 (Web Text))

Max Activating Token Index: 1022

Text #10

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2328. Min Act: -0.0000

Data Index: 362383 (C4 (Web Text))

Max Activating Token Index: 464

Text #11

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2314. Min Act: -0.0000

Data Index: 802543 (C4 (Web Text))

Max Activating Token Index: 799

Text #12

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2160. Min Act: -0.0000

Data Index: 1288956 (C4 (Web Text))

Max Activating Token Index: 104

Text #13

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2203. Min Act: -0.0000

Data Index: 1601687 (Python Code)

Max Activating Token Index: 45

Text #14

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.1917. Min Act: -0.0000

Data Index: 1558766 (Python Code)

Max Activating Token Index: 875

Text #15

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2177. Min Act: -0.0000

Data Index: 679483 (C4 (Web Text))

Max Activating Token Index: 998

Text #16

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2248. Min Act: -0.0000

Data Index: 1560096 (Python Code)

Max Activating Token Index: 408

Text #17

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2160. Min Act: -0.0000

Data Index: 138129 (C4 (Web Text))

Max Activating Token Index: 353

Text #18

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2100. Min Act: -0.0000

Data Index: 674919 (C4 (Web Text))

Max Activating Token Index: 558

Text #19

Max Range: 0.2750. Min Range: -0.2750

Max Act: 0.2154. Min Act: -0.0000

Data Index: 1423351 (Python Code)

Max Activating Token Index: 957