INDEX

Explanations

action before leaving

without loss or optional

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

țiilor

0.55

ালের

0.52

ோம்

0.51

 nếu

0.50

olojik

0.49

 ফস

0.49

 জনে

0.48

াইতে

0.48

ajemy

0.47

atih

0.46

POSITIVE LOGITS

書

0.49

ق

0.43

قاب

0.42

 जमी

0.42

等着

0.41

 Proc

0.40

 cipher

0.39

 überall

0.38

具合

0.38

 दिख

0.38

Activations Density 0.001%