INDEX

Explanations

parts of objects

The neuron fires on two‐word noun phrases that name product components or parts (e.g. “foam base,” “glass top,” “warhead,” “sole grip”).

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

 from

-1.95

in

-1.49

it

-1.48

ığım

-1.42

 voorzien

-1.38

</u>

-1.38

 должен

-1.38

احب

-1.37

 then

-1.35

 vervolgens

-1.34

POSITIVE LOGITS

媵

1.76

 convite

1.60



1.51

蕺

1.49

 lança

1.49

౬

1.45

 alarme

1.40

奩

1.40

 offre

1.40

 example

1.39

Activations Density 0.096%