INDEX

Explanations

conversational interjections

The neuron detects informal conversational markers—especially strings of dots (ellipses), elongated interjections, and filler words like “so,” “but,” or “well.”

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

any

-1.85

任何

-1.82

-1.71

 then

-1.49

 السياسي

-1.38

get

-1.35

買

-1.34

 nostri

-1.34

 that

-1.33

POSITIVE LOGITS

また

1.64

 từ

1.55

since

1.55

 geweldig

1.53

 subido

1.52

および

1.43

 fortfarande

1.42

และ

1.42

béco

1.42

to

1.41

Activations Density 0.018%