INDEX

Explanations

possibility or limitation

could have, could feel

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

 může

0.61

 nidd

0.56

ā

0.56

0.55

 Hopefully

0.53

and

0.52

㪇

0.52

ു

0.49

ittä

0.48

 collabor

0.48

POSITIVE LOGITS

ي

0.74

 ήταν

0.73

י

0.71

だったので

0.61

დან

0.60

이었다

0.60

 있었다

0.59

 лише

0.59

0.57

였다

0.57

Activations Density 0.020%