INDEX

Explanations

section markers like {#sec1-1}

The neuron detects section headings and numbered section identifiers (e.g. “Case Report,” “Materials and Methods,” and their “{#sec…}” tags).

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

 __________

-1.20

 redefined

-1.16

±

-1.16

removing

-1.11

сыз

-1.09

.},

-1.08

-1.07

no

-1.06

 rentrée

-1.06

众人

-1.05

POSITIVE LOGITS

 их

1.43

."

1.38

eloma

1.34

 jejich

1.34

อื่น

1.24

 другим

1.20

 there

1.20

They

1.18

档案馆

1.17

 adonan

1.17

Activations Density 0.147%