INDEX

Explanations

states, conditions, or possessions

This neuron activates on the main article body text (the running content paragraphs), as opposed to titles or section headings.

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

but

-1.36

 πρ

-1.33

 назва

-1.31

少爷

-1.30

ストレッチ

-1.23

といけない

-1.20

=>'

-1.20

rò

-1.19

 别墅

-1.18

 ночи

-1.17

POSITIVE LOGITS

 your

2.52

you

2.02

‌

1.57

isn

1.48

ığınız

1.45

 aren

1.41

}());

1.41

 вам

1.34

おわりに

1.32

You

1.31

Activations Density 0.125%