INDEX

Explanations

academic and scientific terms

The neuron activates on multiword proper‐noun names and technical acronyms (e.g. institutional or program titles, scientific terms, and other named entities).

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

ができます

-1.70

あー

-1.68

Anhalt

-1.49

際には

-1.43

為に

-1.41

Durata

-1.41

あぁ

-1.40

えっ

-1.39

ようになって

-1.37

の方も

-1.37

POSITIVE LOGITS

is

1.52

<bos>

1.46

と

1.41

つまり

1.34

おそらく

1.25

は

1.25

1.23

<sup>

1.23

for

1.23

—

1.22

Activations Density 0.309%