INDEX

Explanations

all followed by a noun or quantifier

The neuron fires on simple plural-copula constructions—i.e. phrases where a plural noun is linked by a form of “to be” (especially “are”) to a predicate.

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

鶫

-1.40

 something

-1.37

鱟

-1.30

 they

-1.26

Alguns

-1.25

つづく

-1.22

 indicando

-1.22

 anything

-1.22

 Dinas

-1.20

 ilang

-1.20

POSITIVE LOGITS

new

1.60

and

1.55

our

1.49

 рестай

1.28

 newly

1.26

1.23

uksen

1.23

 kobieta

1.23

뿐

1.22

 pierwszym

1.20

Activations Density 0.061%