INDEX

Explanations

social media content

The neuron fires on the kind of colloquial, fan-centric “news” language you see in entertainment clickbait headlines—words like “fans,” “leaks,” and informal contractions (apostrophes) that signal hype and reaction.

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

-2.06

 Betrach

-1.51

cimiento

-1.47

érieure

-1.45

 both

-1.41

 Głów

-1.38

 Erfüllung

-1.38

{*}

-1.37

 ﬁrst

-1.37

 prized

-1.34

POSITIVE LOGITS

葙

1.61

 camada

1.54

 TOXIC

1.54

应该

1.53

应该是

1.51

他们

1.49

 LIABLE

1.48

whom

1.47

alucía

1.47

 باید

1.43

Activations Density 0.121%