INDEX

Explanations

the following technical word

The neuron detects structural labels that introduce embedded content blocks—words like “table,” “video,” “code,” “function,” or “option” that head examples, listings, or sections.

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

 with

-1.82

in

-1.59

as

-1.55

 titoli

-1.48

 obicei

-1.48

 apsau

-1.47

 belangrijk

-1.47

 duomen

-1.46

 ugyan

-1.45

POSITIVE LOGITS

1.91

1.90

1.89

1.81

1.74

The

1.73

1.71

 elusive

1.65

н

1.50

1.48

Activations Density 0.101%