INDEX

Explanations

content version format

The neuron triggers on file‐format and metadata/control tokens (e.g. RTF tags, HTML/XML meta headers, PO/POT markers, encoding declarations)—i.e. on document‐formatting or metadata directives.

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

Qar

-0.79

 acoper

-0.79

حه

-0.78

 対策

-0.76

 sixty

-0.75

papel

-0.73

 forty

-0.73

 mijne

-0.72

roughts

-0.71

 aquellos

-0.70

POSITIVE LOGITS

 version

1.14

begin

1.14

version

1.03

Version

0.96

стема

0.93

 Version

0.92

 versions

0.91

 версия

0.91

 units

0.89

 Pilihan

0.88

Activations Density 0.004%