INDEX

Explanations

array followed by brackets

This neuron detects ampersand characters (e.g. “&” and HTML entity starts like “&frac;”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 beginning

0.76

 beginnings

0.68

beginning

0.64

 begin

0.59

lac

0.57

onym

0.55

 آغاز

0.54

 really

0.54

വരുടെ

0.53

 початку

0.53

POSITIVE LOGITS

ﺹ

0.63

umlu

0.56

equivariant

0.55

}`);

0.54

 معروف

0.54

 মালা

0.53

 مدارس

0.53

 एससी

0.52

 بود

0.52

শোধ

0.51

Activations Density 0.014%