INDEX

Explanations

distinctive personal characteristics

The neuron fires on terms emphasizing uniqueness or individual/brand identity—words like “identity,” “distinct,” “personality,” “fingerprints,” and “mark.”

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

-1.79

↵↵

-1.59

the

-1.57

</i>

-1.46

]:

-1.45

us

-1.39

):

-1.38

 amplio

-1.36

</sub>

-1.35

 them

-1.34

POSITIVE LOGITS

৹

1.66

وانید

1.45

1.43

 skola

1.43

 preocupado

1.43

 навіть

1.42

 kartą

1.41

 bazen

1.39

&&

1.38

以上です

1.36

Activations Density 0.043%