INDEX

Explanations

what does it mean?

This neuron detects occurrences of the “mean”/“matter” lexeme (e.g. English “mean” in “what does … mean?” or Russian “дело” in “в чем дело?”) in questions.

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

to

-1.34

But

-1.22

probably

-1.16

zaSy

-1.16

The

-1.14

of

-1.13

 your

-1.11

>?

-1.09

marzo

-1.09

 wesentlich

-1.04

POSITIVE LOGITS

3.47

?”

1.44

domada

1.41

もありますが

1.39

 Préférences

1.37

Preparazione

1.30

kulum

1.28

 بیمارستان

1.28

もありました

1.27

不过是

1.27

Activations Density 0.064%