INDEX

Explanations

how

The neuron primarily activates on numeric tokens—especially standalone numbers or decimal measurements.

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

د

0.80

ஒரு

0.79

دخ

0.70

्

0.69

এক

0.68

↵

0.68

س

0.68

The

0.68

એ

0.68

に

0.67

POSITIVE LOGITS

neapolis

0.87

 Обще

0.84

 Nauchno

0.84

Ⴎ

0.83

 novem

0.81

 januar

0.81

habalar

0.79

 onderwerp

0.79

aas

0.79

రుగులు

0.77

Activations Density 0.001%