INDEX

Explanations

determiner followed by a noun or adjective

This neuron is primarily responding to the English articles “a” and “the” when they head noun 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

 which

-1.60

 using

-1.48

нача

-1.46

 ihrer

-1.30

 including

-1.30

 поставки

-1.29

 которые

-1.27

<>());

-1.25

 they

-1.25

ії

-1.23

POSITIVE LOGITS

も同じ

1.45

 těchto

1.44

 anahtar

1.39

描く

1.37

 wtedy

1.36

に似

1.36

 тщательно

1.34

 fiyatı

1.32

を探す

1.32

<td>

1.31

Activations Density 0.146%