INDEX

Explanations

our decision or feeling

structured text and code elements

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 siren

0.52

ILLS

0.47

 тоді

0.45

 grenade

0.45

 зада

0.44

 burdensome

0.44

LLS

0.44

யி

0.43

 bathrobe

0.43

соеди

0.43

POSITIVE LOGITS

تر

0.48

fassung

0.47

sprach

0.45

Matches

0.45

 especializado

0.45

 março

0.44

 Recursos

0.44

اگر

0.44

 kalau

0.43

 formado

0.43

Activations Density 0.001%