INDEX

Explanations

criteria or option selection

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 khóa

-0.81

skins

-0.79

ỳnh

-0.79

のではない

-0.79

candidate

-0.79

 Further

-0.78

 znaczy

-0.77

 Already

-0.76

führ

-0.76

Hö

-0.75

POSITIVE LOGITS

 crit

1.03

Crit

0.96

 criterios

0.94



0.93

 criterion

0.91

 Crit

0.91

 criteria

0.89

cri

0.88

 Criteria

0.88

 criter

0.84

Activations Density 0.001%