INDEX

Explanations

two

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dieux

-0.69

 vägen

-0.65

 oreilles

-0.64

 lèvres

-0.63

 preuves

-0.63

 écrits

-0.62

 épaules

-0.61

 nuages

-0.60

 operativos

-0.60

the

-0.60

POSITIVE LOGITS

 nakalista

0.84

 thirds

0.81

ArrowToggle

0.81

 continents

0.79

 wrongs

0.79

 kinds

0.74

 distinct

0.73

AccessorTable

0.73

 guineas

0.73

 batches

0.73

Activations Density 0.027%