INDEX

Explanations

studies of relation between concepts

This neuron detects the word “of” when it appears in structured scientific‐study or paper‐title phrases (e.g. “Role of,” “Study of,” “Application of”).

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

 that

-1.53

one

-1.49

 until

-1.37

-1.34

 earlier

-1.33

 before

-1.31

 many

-1.31

-1.30

 вашим

-1.30

 first

-1.28

POSITIVE LOGITS

quelles

1.38

 hogyan

1.31

Пото

1.28

 milyen

1.27

这个问题

1.25

㈀

1.25

diven

1.24

ఁ

1.24

应该

1.23

quels

1.23

Activations Density 0.382%