INDEX

Explanations

removal of

designing, extracting, questions

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 tří

0.54

 nanny

0.51

 každý

0.48

 volleyball

0.48

RootManager

0.48

Ú

0.47

 achie

0.47

 Gómez

0.47

 romant

0.47

 Raila

0.47

POSITIVE LOGITS

]:

0.47

意思

0.47

敒

0.46

<0x0D>

0.46

提示

0.44

designs

0.44

ینی

0.44

Designing

0.43

奈良

0.42

Pennsylvania

0.42

Activations Density 0.000%