INDEX

Explanations

the idiom

The neuron flags content words that appear in common English idioms (e.g. plate, curve, bandwagon, fire, party).

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

 rése

-1.39

 tenet

-1.37

殽

-1.37

Comes

-1.37

⿸

-1.35

tifies

-1.34

또

-1.33

the

-1.33

莅

-1.31

 triko

-1.30

POSITIVE LOGITS

 didn

1.75

."

1.74

was

1.70

__':

1.68

',

1.63

 doesn

1.60

</em>

1.59

')

1.59

 become

1.59

Activations Density 0.107%