INDEX

Explanations

constantly

The neuron strongly activates on content‐bearing, low‐frequency or domain‐specific words (technical terms, proper nouns, and uncommon adjectives/nouns) 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

 produse

0.71

 produkter

0.70

 AFTER

0.70

 rehearing

0.70

 Ergebnisse

0.67

 risultati

0.67

 jeżeli

0.67

 producten

0.67

 resultados

0.66

 eredmény

0.66

POSITIVE LOGITS

 overarching

1.06

 bustling

0.94

 khắp

0.93

 boyunca

0.84

 patrolling

0.83

🫶

0.79

 κάθε

0.78

 bốn

0.77

 🥰

0.77

 intricately

0.77

Activations Density 0.437%