INDEX

Explanations

programming, licensing, summary, media

The neuron activates on sequences of digits (numeric literals), particularly multi‐digit numbers (e.g. version numbers, years, counts) in technical 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

 only

-1.10

 their

-1.03

 truly

-0.86

 vrst

-0.86

 required

-0.86

 Política

-0.86

 while

-0.84

its

-0.84

 APPELL

-0.83

 requires

-0.83

POSITIVE LOGITS

junit

0.98

haping

0.95

)}(\

0.93

knew

0.88

vlog

0.87

org

0.87

 heridos

0.86

ProductCategory

0.85

 empez

0.83

 junit

0.82

Activations Density 0.000%