INDEX

Explanations

mathematical symbols

The neuron is detecting LaTeX‐style math notation (backslash commands, superscripts, primes, etc.) in the text.

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

by

-1.52

などを

-1.37

が

-1.36

to

-1.30

 have

-1.27

 also

-1.27

是被

-1.26

などが

-1.23

と

-1.23

で

-1.19

POSITIVE LOGITS

 Outils

1.39

 italie

1.37

ܖ

1.34

 balení

1.33

vaný

1.32

vemos

1.31

場合には

1.31

 szob

1.30

 Pièces

1.29

んだよね

1.29

Activations Density 0.025%