INDEX

Explanations

apologies and regret

The neuron flags expressions of apology or regret (e.g. “sorry,” “feel bad/terrible,” “pity party”).

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

後記

-1.68

掼

-1.61

欷

-1.59

小伙伴

-1.59

 biß

-1.58

 utilizó

-1.57

 that

-1.55

 erhältlich

-1.55

 ricerca

-1.54

 deft

-1.53

POSITIVE LOGITS

Même

1.58

»;

1.57

).

1.57

」

1.54

 Enciclopedia

1.52

 cadena

1.48

 Belén

1.48

 Republica

1.43

 capitan

1.43

 latino

1.39

Activations Density 0.001%