INDEX

Explanations

Seo Optimization, Rapid changes, matching records

The neuron fires strongly on numeric tokens (digits, multi-digit sequences, version numbers) in the text.

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

 boulangerie

-0.77

ymas

-0.74

 degust

-0.73

 Synthesis

-0.71

 چای

-0.71

 Vasili

-0.70

 Basis

-0.69

Moonlight

-0.69

zada

-0.69

RTL

-0.69

POSITIVE LOGITS

cr

0.91

krish

0.90

 dominating

0.88

 companies

0.84

 company

0.84

 crutches

0.84

cr

0.83

 SUPER

0.82

Cr

0.81

 context

0.81

Activations Density 0.007%