INDEX

Explanations

wondering

The neuron predominantly fires on discourse‐structuring function words—especially subordinating/connective words like “about,” “how,” “that,” and similar markers that introduce clauses.

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

ньо

0.77

 σας

0.67

כם

0.66

 enterprising

0.65

 বলুন

0.64

 gamers

0.63

 firms

0.62

ですね

0.60

ครับ

0.60

একটি

0.60

POSITIVE LOGITS

ตัวเอง

0.93

 injust

0.91

ตน

0.90

 Myself

0.86

 injustice

0.83

 خودم

0.83

 helplessness

0.82

 myself

0.80

 своему

0.79

 아버지

0.79

Activations Density 0.349%