specific entities and their context

The neuron fires on domain-specific proper names and technical terminology (e.g. brand names, scientific terms, and specialized jargon).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

razer

-1.05

🩻

-1.02

の人が

-0.99

wachsene

-0.99

nissan

-0.99

 tarifas

-0.99

sniper

-0.97

Isi

-0.96

phim

-0.95

𓆸

-0.94

POSITIVE LOGITS

ฯ

1.38

0.96

0.93

0.92

0.91

 различные

0.90

uk

0.89

 dioxide

0.89

Activations Density 0.325%