INDEX

Explanations

names and titles

The neuron flags tokens that are part of named entities—primarily proper nouns (people’s names, organizations, titles, etc.).

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

and

-2.86

-1.77

 after

-1.55

now

-1.46

宴会

-1.40

</i>

-1.38

 ahora

-1.38

 не

-1.37

してます

-1.37

 mientras

-1.35

POSITIVE LOGITS

1.88

殽

1.88

敺

1.79

 иной

1.73

 ideen

1.72

 universitario

1.70

 ordinario

1.68

瑭

1.67

⮑

1.65

lizes

1.63

Activations Density 0.001%