INDEX

Explanations

when a or an

The neuron is detecting occurrences of common clause- or sentence-introducing function words—especially the indefinite articles “a” and “an” (and similar tokens like “When”) at the starts of phrases.

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

是怎么回事

-1.43

考えた

-1.32

 možnost

-1.27

可能有

-1.24

 Problemas

-1.23

感じた

-1.20

 doctrina

-1.19

 prazo

-1.19

ремя

-1.19

 Aunque

-1.18

POSITIVE LOGITS

 this

1.60

 different

1.60

of

1.48

 pirm

1.45

gucigalpa

1.44

 into

1.43

 її

1.27

 appunto

1.27

 tiek

1.27

 sentra

1.24

Activations Density 0.188%