INDEX

Explanations

client/server architecture

The neuron is tuned to pick out relatively rare, content-bearing tokens (e.g. specific numbers, proper nouns, and uncommon nouns).

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

punkt

0.73

anza

0.72

эф

0.71

 profil

0.70

 صلاحیت

0.70

 wnios

0.70

0.69

 governance

0.69

gge

0.69

 rationale

0.68

POSITIVE LOGITS

 Bitmap

0.78

 Feuilles

0.76

 Memories

0.76

JAVA

0.74

 някои

0.74

 Beberapa

0.73

 Eyes

0.73

 Arten

0.73

 Discounts

0.73

一些

0.72

Activations Density 0.000%