INDEX

Explanations

parts of words

The neuron fires on longer, meaningful subword fragments (morphemes or affixes) within words—e.g. “ban,” “trick,” “ling,” or “cross.”

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

 خدمة

-1.30

-1.21

——–

-1.21

าหลี

-1.20

 Know

-1.18

"_"

-1.18

 Viele

-1.17

 있으며

-1.16

ÄR

-1.13

 vielen

-1.12

POSITIVE LOGITS

 montaż

1.42

昔から

1.38

 soirée

1.37

晅

1.37

 with

1.37

χύ

1.29

佧

1.27

ͺ

1.24

cję

1.23

[]):

1.22

Activations Density 0.090%