INDEX

Explanations

junit tests

the word "junit" in Java import statements and 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

 That

-3.22

 Regarding

-3.17

 When

-3.00

 Roughly

-2.98

 Specifically

-2.86

 Actually

-2.78

 Certainly

-2.73

 Interestingly

-2.73

 Compared

-2.69

 Importantly

-2.69

POSITIVE LOGITS

塒

2.66

ar

2.61

in

2.45

蚡

2.22

Ꮓ

2.22

2.20

 kere

2.17

蹕

2.11

ila

2.09

身旁

2.06

Activations Density 0.003%