INDEX

Explanations

outcome or cause

The neuron fires on salient content words—nouns, verbs, and adjectives carrying the main informational meaning of a sentence.

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

or

-1.65

Koordinaten

-1.59

as

-1.52

</th>

-1.48

 hodně

-1.48

 затем

-1.45

and

-1.38

 欣赏

-1.38

ayaquil

-1.37

八年

-1.37

POSITIVE LOGITS

despite

1.48

1.42

 pertinente

1.42

١

1.41

ほんと

1.40

スムーズ

1.38

ちょこ

1.37

摸了

1.35

ようになります

1.34

やっぱ

1.34

Activations Density 0.063%