INDEX

Explanations

code block terminators

the presence of numeric literals (number constants) in code.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 ponad

-0.93

Estimated

-0.86

 such

-0.85

 wpływ

-0.82

 Pflicht

-0.81

 wcześ

-0.81

 Earlier

-0.80

Şi

-0.80

숲

-0.79

Earlier

-0.78

POSITIVE LOGITS

helm

0.82

 男子

0.77

 način

0.74

 sherds

0.71

onnage

0.71

ufc

0.71

rte

0.70

Ory

0.70

とは言

0.69

Activations Density 0.010%