INDEX

Explanations

codes and questions

This neuron detects code format specifiers (e.g. the tokens “html” and “js”) indicating response‐format declarations in code.

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

⿻

-1.52

事に

-1.50

様な

-1.49

に入れる

-1.46

書の

-1.38

為に

-1.38

 naprawdę

-1.37

に入る

-1.35

 Tento

-1.33

ところに

-1.32

POSITIVE LOGITS

 only

1.43

を

1.41

<bos>

1.40

は

1.34

but

1.34

や

1.32

if

1.29

！！！！

1.25

 generally

1.23

が

1.21

Activations Density 0.000%