INDEX

Explanations

algebraic expressions with j

The neuron activates on uppercase technical acronyms and language/standard names (e.g. VHDL, IEEE, Ada).

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

 zaczą

-1.45

of

-1.38

-1.36

㗎

-1.35

our

-1.34

 szuka

-1.31

 penghargaan

-1.31

搵

-1.28

很難

-1.28

POSITIVE LOGITS

anow

1.54

 Anliegen

1.39

 bons

1.34

зраст

1.30

 know

1.29

medriver

1.29

ментов

1.27

自分の

1.27

adeira

1.27

除了

1.26

Activations Density 0.017%