INDEX

Explanations

introducing clauses

The neuron fires on words that open new sentences or clauses—i.e. sentence-initial discourse markers and conjunctions (To, After, If, As, With, Under, This, etc.).

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

})$

-1.34

Напомним

-1.27

Bahkan

-1.27

 vermek

-1.27

labilir

-1.26

-1.24

 étan

-1.24

/>

-1.23

Demikian

-1.22

 mejores

-1.20

POSITIVE LOGITS

 jego

1.39

its

1.39

 there

1.38

 using

1.37

 only

1.30

 anstatt

1.21

 atriz

1.20

 seiner

1.16

in

1.14

 involved

1.13

Activations Density 0.045%