INDEX

Explanations

mathematical symbols and abbreviations

The neuron fires on tokens that are escape‐ or metacharacter punctuation (especially backslashes “\” and similar regex/LaTeX delimiters).

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

 which

-2.02

but

-1.61

だが

-1.30

Sá

-1.27

⌒

-1.23

ства

-1.23

rän

-1.21

Pâ

-1.20

mények

-1.20

 milla

-1.19

POSITIVE LOGITS

1.49

 récents

1.45

besos

1.41

現在の

1.41

 récentes

1.38

jugador

1.36

 adatok

1.34

 zwyk

1.34

}}=$

1.31

—

1.29

Activations Density 0.122%