INDEX

Explanations

words ending in s, d, or y

The neuron selectively activates on subword pieces that occur at the ends of words (i.e. word-ending or suffix fragments).

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

</strong>

-1.53

 estrenar

-1.47

 lohnt

-1.45

You

-1.40

uacute

-1.39

 всегда

-1.35

 одной

-1.34

 capabilities

-1.33

crania

-1.29

 nekaj

-1.29

POSITIVE LOGITS

 sosis

1.45

 Souhaitez

1.30

 then

1.29

 itinerario

1.26

 kasar

1.23

就已经

1.22

 rajut

1.22

糶

1.20

1.19

 susun

1.19

Activations Density 0.456%