INDEX

Explanations

annual meetings and conferences

The neuron detects references to academic or professional conference events—especially “Annual Meeting”‐style names, dates, and related numeric tokens.

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

 broth

-0.78

fashion

-0.77

opers

-0.77

ージュ

-0.75

 necessary

-0.73

/?

-0.70

chowa

-0.70

 permitido

-0.70

uge

-0.69

 filling

-0.69

POSITIVE LOGITS

 meeting

1.74

 meetings

1.57

Meeting

1.49

 Meetings

1.45

meeting

1.37

 Meeting

1.37

Meetings

1.31

 annual

1.20

 MEETING

1.16

meetings

1.16

Activations Density 0.024%