INDEX

Explanations

partner id

This neuron activates on non-Latin script tokens (e.g. Chinese characters or other CJK/Cyrillic/Japanese tokens).

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

the

-3.38

</b>

-2.86

and

-2.64

 ...,

-2.50

𐑂

-2.41

 попыта

-2.36

-2.34

should

-2.33

và

-2.27

あぁ

-2.23

POSITIVE LOGITS

เกง

2.30

ции

2.27

мені

2.08

2.05

zystać

2.05

Отзывы

2.02

י

2.00



1.98

ſelves

1.98

水墨

1.97

Activations Density 0.001%