INDEX

Explanations

common prefixes

prefixes like non-, multi-, pre-, post-, anti-

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

Კ

0.38

 möglicherweise

0.35

ോട്ട്

0.35

 হইয়৷

0.34

}.

0.34

 wurden

0.34

 kommt

0.33

 maken

0.33

 ########.

0.33

 ikke

0.33

POSITIVE LOGITS

0.86

‐

0.64

0.57

־

0.55

0.52

‑

0.47



0.44

¬

0.44

-‘

0.41

0.40

Activations Density 0.397%