INDEX

Explanations

foreign language words

This neuron is highly activated by informal discourse markers and hesitation sounds—i.e. conversational “fillers” like “ahem,” “uh,” and “er.”

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

 even

-1.97

-1.72

</h4>

-1.70

 with

-1.64

 only

-1.51

of

-1.50

-1.42

in

-1.40

-1.35

POSITIVE LOGITS

︓

1.83

Чита

1.63

 restoran

1.60

lijden

1.59

柳

1.59

 russo

1.56

 kriti

1.56

 우리

1.55

unser

1.55

 occident

1.52

Activations Density 0.043%