INDEX

Explanations

relates to

The neuron fires on formal technical/patent-style wording—especially “invention,” “relates,” “present invention,” section headings, and other legal-technical jargon.

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

if

-1.69

-1.41

of

-1.34

 they

-1.32

}$.

-1.30

 несмотря

-1.30

alignSelf

-1.26

the

-1.23

 weshalb

-1.22

</i>

-1.20

POSITIVE LOGITS

靦

1.50

今は

1.49

簡

1.38

 construida

1.38

揾

1.35

 battre

1.35

今年は

1.31

 regardé

1.30

 diseñada

1.30

 推薦

1.30

Activations Density 0.205%