INDEX

Explanations

assigning numeric codes

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

quilibrium

-0.84

elock

-0.80

 betrachtet

-0.78

 khối

-0.74

 الماضي

-0.73

 deque

-0.73

AccessControl

-0.73

 título

-0.72

териал

-0.71

ēm

-0.71

POSITIVE LOGITS

 coding

3.30

Coding

2.72

 Coding

2.72

coding

2.58

 coded

2.48

 coder

2.08

coded

2.08

 codes

2.06

 code

1.91

 scoring

1.84

Activations Density 0.091%