INDEX

Explanations

celestial objects, weather, contrasting ideas

The neuron strongly activates on “can be” statements—i.e. passive‐voice modal constructions indicating something “can be” done (e.g. “can be transferred,” “can be seen”).

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

0.88

 brainstorming

0.79

 brainstorm

0.77

two

0.74

 five

0.74

 identify

0.74

 informal

0.73

 teaspoon

0.73

 prerequisite

0.73

 tabletop

0.73

POSITIVE LOGITS

的存在

0.80

 ਨਹੀਂ

0.75

angulo

0.74

jwt

0.73

сама

0.73

expires

0.72

となりました

0.71

ēja

0.71

闇

0.71

urgia

0.71

Activations Density 0.000%