INDEX

Explanations

actions and invariants

The neuron fires primarily on the word “action” (or “actions”), especially in mathematical contexts.

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

formis

-1.49

е

-1.45

 duração

-1.44

 kurzem

-1.43

–

-1.43

 déclaration

-1.39

༛

-1.37

end

-1.34

In

-1.34

 consciência

-1.31

POSITIVE LOGITS

 Scritto

1.48

⟝

1.45

 armi

1.40

 gioia

1.36

AfterViewInit

1.35

は

1.34

ngilizce

1.32

Odpowiedz

1.30

 概要

1.30

chowa

1.30

Activations Density 0.016%