INDEX

Explanations

introduces examples or features

This neuron detects discourse markers that introduce highlights or list items—words like “Another,” “Famous,” “Highlight(ing),” or “One” at the start of a new point.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

主に

0.58

 sekä

0.58

 samt

0.57

めっちゃ

0.55

 うん

0.55

+"\

0.55

全て

0.54

blah

0.54

 blah

0.53

 ใช่

0.52

POSITIVE LOGITS

的是

1.87

 adalah

1.66

 ialah

1.63

are

1.59

is

1.55

 jsou

1.55

 आहेत

1.51

 হলেন

1.51

was

1.50

是在

1.50

Activations Density 0.609%