INDEX

Explanations

city names

This neuron activates on geographic place names (cities, states, countries) in document metadata.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

of

-1.55

 nella

-1.38

we

-1.34

 innerhalb

-1.33

})}

-1.32

herited

-1.32

 have

-1.31

our

-1.30

 molti

-1.27

compared

-1.27

POSITIVE LOGITS

 Initially

1.44

 usually

1.41

 Usually

1.40

strator

1.35

 Often

1.32

 Shortly

1.31

Usually

1.30

还真是

1.29

Often

1.28

着头

1.27

Activations Density 0.106%