INDEX

Explanations

modified wood / focused response

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

evice

1.90

ke

1.67

kého

1.60

𝐡

1.60

giers

1.58

oš

1.56

hidupan

1.56

uhan

1.55

lesque

1.54

𝐝

1.54

POSITIVE LOGITS

 déterminer

1.74

\|

1.73

 QUANT

1.69

 конкре

1.67

 pointers

1.66

在哪

1.64

 откуда

1.64

Predicate

1.64

 acert

1.60

 αγο

1.60

Activations Density 0.738%