INDEX

Explanations

prefix, suffix, and title markers

The neuron strongly activates on terms that name or denote positional string or name segments—e.g. “prefix,” “middle,” “initial,” “suffix,” or “postfix.”

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

 only

-2.08

<bos>

-1.63

usw

-1.40

 тільки

-1.38

鱟

-1.29

 primas

-1.28

-1.27

 mercato

-1.27

 anum

-1.26

 even

-1.23

POSITIVE LOGITS

 parfois

1.45

 eachother

1.33

 muitas

1.33

 muitos

1.30

 vielen

1.30

aimeJ

1.29

 بعد

1.28

 ซึ่ง

1.27

megen

1.24

 nombreuses

1.23

Activations Density 0.041%