INDEX

Explanations

initials followed by names

The neuron consistently fires on instances of capital‐letter initials followed by a period (as in “J.” or “B.”) within names or citations.

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

 förstå

-1.52

]$}

-1.45

 svårt

-1.42

Neden

-1.33

 guerrero

-1.30

الدرا

-1.30

比较好

-1.28

Ꮸ

-1.28

Adapun

-1.27

Berdasarkan

-1.27

POSITIVE LOGITS

and

1.66

</h2>

1.65

 Percent

1.46

</em>

1.35

Ḧ

1.33

⮚

1.24

 vroeger

1.22

</h3>

1.21

 créée

1.20

 fracas

1.19

Activations Density 0.035%