INDEX

Explanations

mathematical descriptors and structures

The neuron detects descriptive mathematical modifiers (adjectival qualifiers) attached to concepts (e.g. “compact,” “contractive,” “scaled,” “finite,” “nonempty,” “monic,” etc.).

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

 maillot

-1.33

деров

-1.32

 coupé

-1.31

鮰

-1.31

 diamant

-1.31

 extraordin

-1.30

 inclus

-1.30

 formid

-1.27

 aggres

-1.26

 ۱۰

-1.25

POSITIVE LOGITS

 acogedor

1.36

 there

1.31

 keterampilan

1.30

 красивая

1.27

 дори

1.23

 всеки

1.23

There

1.23

 frecuentemente

1.20

人是

1.20

komendasi

1.20

Activations Density 0.056%