INDEX

Explanations

strong emotional reactions

impressed by

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

 knowing

0.65

 having

0.62

knowing

0.60

having

0.55

 thoughts

0.55

 imagining

0.53

Knowing

0.52

 feelings

0.50

 getting

0.50

 being

0.49

POSITIVE LOGITS

 насколько

0.58

 spezif

0.46

 сколько

0.46

 wymaga

0.45

 органами

0.45

 специфи

0.45

iede

0.43

meniz

0.43

 cụ

0.42

pdb

0.42

Activations Density 0.031%