INDEX

Explanations

against or overcome

The neuron fires on constructions where a verb and its object are followed by the preposition “for” (e.g. “no match for,” “stop them for,” “cannot withstand for”), i.e. verb-object-for patterns.

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

-1.79

</h2>

-1.60

 verbess

-1.44

 häufiger

-1.32

 verstär

-1.30

 других

-1.27

 öfter

-1.26

 other

-1.22

 anything

-1.20

Beschrijving

-1.20

POSITIVE LOGITS

ᾰ

1.45

 Méd

1.38

ようやく

1.35

鱚

1.30

 frère

1.30

 laki

1.30

橚

1.29

 that

1.28

 takimi

1.28

开心

1.27

Activations Density 0.028%