INDEX

Explanations

proper nouns after English translation

This neuron detects the punctuation and quotation‐mark tokens used to introduce or enclose parenthetical translations or literal‐translation labels (e.g. the “: ” and opening quote around translated terms).

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

你是不是

-1.59

ubahan

-1.52

叭

-1.51

しているので

-1.46

你先

-1.45

 Thats

-1.45

仕事の

-1.45

 некоторых

-1.41

卧槽

-1.41

 sosteniendo

-1.38

POSITIVE LOGITS

)”.

1.52

 although

1.49

 that

1.46

as

1.45

<td>

1.39

but

1.38

”…

1.38

⃣

1.36

💗

1.35

 despite

1.34

Activations Density 0.006%