INDEX

Explanations

problems or negative states

code and abstract concepts

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

 credible

0.47

 detectable

0.44

 recognizable

0.44

 adapté

0.43

ቸው

0.43

 recognisable

0.42

 liberté

0.42

 enforceable

0.42

 свобо

0.41

recognizable

0.40

POSITIVE LOGITS

 existente

0.57

 otherwise

0.51

 vorhand

0.48

 existentes

0.47

existing

0.46

otherwise

0.45

 থাকলে

0.45

ያ

0.45

Existing

0.44

 путем

0.44

Activations Density 0.468%