INDEX

Explanations

many types of people or things

The neuron fires on the head noun of broad “generalizing” statements introduced by quantifiers (e.g. “Many …,” “Most …”).

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

 花纹

-1.50

 soooo

-1.48

 素描

-1.45

粜

-1.41

 allerlei

-1.41

 sooo

-1.39

 内衣

-1.39

 手绘

-1.38

 estadounid

-1.38

 vraiment

-1.38

POSITIVE LOGITS

 will

2.50

 have

1.71

of

1.63

\}$.

1.50

 donnera

1.49

 ended

1.48

 enables

1.48

 politiet

1.47

 changed

1.45

 change

1.45

Activations Density 0.107%