INDEX

Explanations

specific

The neuron fires on tokens that signal precision or specification—words like “exactly,” “specific,” “unique,” “virtually,” “with,” and “to” that typically appear when something is being narrowly defined or qualified.

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

 keduanya

0.69

两者

0.67

 both

0.65

 semuanya

0.64

Both

0.61

both

0.60

ពេល

0.59

の両

0.58

Including

0.55

दोनों

0.55

POSITIVE LOGITS

特定的

2.92

特定の

2.88

特定

2.83

 tertentu

2.72

 specific

2.69

 특정

2.67

 συγκεκρι

2.63

 particular

2.50

 特定

2.50

specific

2.39

Activations Density 1.107%