INDEX

Explanations

honorifics followed by names

The neuron activates on uppercase abbreviations or honorifics that include periods (e.g. “U.S.,” “Sen.,” “Col.,” “Sgt.”).

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

或

-1.86

There

-1.77

Sometimes

-1.70

If

-1.65

Even

-1.64

 или

-1.57

All

-1.55

Often

-1.55

Many

-1.54

POSITIVE LOGITS

who

1.88

mpagne

1.66



1.54

渼

1.53



1.51

 dança

1.51

踵

1.49

鮰

1.48

tempts

1.48

簠

1.47

Activations Density 0.009%