INDEX

Explanations

special characters and punctuation

This neuron is keyed to tokens that are formatting or metadata markers—long runs of punctuation, numbers, or separator-style lines (page numbers, horizontal bars, hashes, etc.).

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

 $^{\

-1.98



-1.97

ra

-1.95

.''

-1.95

,\\

-1.93

 devem

-1.92

pecabe

-1.91

</em>

-1.91

-\\

-1.89

 possu

-1.87

POSITIVE LOGITS

</h1>

2.08

 Jér

1.94

 یا

1.88

𐐪

1.85

 retratos

1.85

тории

1.81

すぐに

1.79

中です

1.78

 decoradas

1.77

～♪

1.73

Activations Density 0.026%