INDEX

Explanations

whenever I explain

This neuron activates on domain-specific technical terms—especially mathematical or chemical jargon (like “number,” “isomers,” “integrating,” “inner product,” 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

 관리

-1.15

:^{

-1.13

Firma

-1.13

 adatok

-1.13

 تعريف

-1.11

 хай

-1.11

に対する

-1.09

DisplayMetrics

-1.08

 السياس

-1.06

 niebies

-1.06

POSITIVE LOGITS

 usually

1.81

 always

1.77

 often

1.71

 sometimes

1.58

 ALWAYS

1.55

 their

1.41

 всегда

1.41

֩

1.39

 each

1.38

 Sometimes

1.38

Activations Density 0.093%