INDEX

Explanations

Rudy followed by a name

The neuron strongly activates on capitalized named entities and acronyms (proper nouns and multi‐letter uppercase tokens).

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

 Zeitraum

-2.92

 byproduct

-2.88

’

-2.84

 berühmten

-2.83

 zweiten

-2.56

 ewigen

-2.50

 mahdol

-2.47

 bemerkt

-2.42

 Gegensatz

-2.41

 täglichen

-2.41

POSITIVE LOGITS

sè

2.84

栳

2.72

2.70

2.67

oe

2.55

真的好

2.50

缡

2.47

⃢

2.41

僰

2.36

鶿

2.36

Activations Density 0.007%