INDEX

Explanations

famous movie and myth characters

The neuron is detecting capitalized proper nouns—especially names of people, characters, and titled works.

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

 męski

-1.39

";

-1.36

};

-1.35

 Preußen

-1.34

):

-1.33

 them

-1.33

--

-1.25

is

-1.24

 pudieron

-1.23

 moaning

-1.23

POSITIVE LOGITS

 prestigio

1.70

 mando

1.42

abancı

1.42

䊉

1.42

但却

1.41

 schutz

1.40

 azulejos

1.36

 motore

1.35

OUGHT

1.34

薅

1.34

Activations Density 0.018%