INDEX

Explanations

The neuron strongly fires on formatted headings or section titles—i.e. phrases set off by colons or em-dashes (for example “Job Fair – Round One” or “Last week: …”).

days and times

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

who

-1.11

 acestei

-1.05

behälter

-1.00

-0.99

 additions

-0.97

<em>

-0.96

ep

-0.96

||

-0.95

POSITIVE LOGITS

Če

1.26

Ruj

1.13

Παρα

1.09

Čes

1.09

):}

1.05

Več

1.05

ваете

1.05

Karakteristik

1.03

Familien

1.02

 schreck

1.02

Activations Density 0.030%