INDEX

Explanations

entities and entity classification

The neuron fires on relatively rare or specialized content words—especially technical or domain‐specific nouns and adjectives—rather than on common function words.

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

 earthquakes

0.61

 codons

0.59

 యుద్ధ

0.58

 トート

0.58

 goles

0.56

 конференции

0.55

 songs

0.53

 thunderstorms

0.52

 vét

0.52

 साड़ी

0.51

POSITIVE LOGITS

 entities

1.60

 entity

1.52

 devices

1.35

entities

1.34

 Entities

1.34

 dispositivo

1.27

 zariaden

1.27

 device

1.25

アイテム

1.24

entity

1.23

Activations Density 1.279%