INDEX

Explanations

for, to, understanding

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

が見える

-1.00

涝

-0.93

рованная

-0.86

ğinde

-0.85

 tilby

-0.84

Steg

-0.82

くなって

-0.81

tead

-0.81

 hjemmeside

-0.81

tira

-0.81

POSITIVE LOGITS

 understanding

2.83

 determining

2.50

 analysis

2.45

 studying

2.38

 assessing

2.36

 investigating

2.31

 study

2.27

 determine

2.25

 evaluating

2.25

分析

2.23

Activations Density 0.423%