INDEX

Explanations

separately

The neuron is a language detector that fires on Cyrillic‐script (Russian) 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

↵↵

-1.38

 birkaç

-1.36

’，

-1.32

 impoverished

-1.31

诸多

-1.27

 decipher

-1.24

咻

-1.20

 paramount

-1.18

 JUST

-1.16

 countless

-1.16

POSITIVE LOGITS

 نیز

1.71

蹀

1.50

蚡

1.47

⑼

1.41

[:,

1.38

Ꮓ

1.38

کیف

1.37

훙

1.35

 niveles

1.34

着她

1.31

Activations Density 0.001%