INDEX

Explanations

initials followed by a name

The neuron strongly activates on isolated uppercase single‐letter tokens followed by a period (e.g. “J.”, “C.”, “D.”) used as initials or abbreviations.

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

-1.71

Esperamos

-1.51

Olvid

-1.48

Aún

-1.45

 vostra

-1.44

 pido

-1.41

 aconte

-1.39

Seguir

-1.37

lusst

-1.35

 suoi

-1.34

POSITIVE LOGITS

to

1.87

]}$

1.56

(«

1.52

de

1.51

 (‘

1.48

 (“

1.47

 ‘’

1.41

帷

1.40

1.38

be

1.38

Activations Density 0.043%