INDEX

Explanations

modelhas several variations

The neuron responds to numeric and ordinal expressions—years, model numbers, “first,” “second,” and similar quantified tokens.

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

DummyView

0.61

Fragmentation

0.57

medieval

0.57

❚

0.56

汽车

0.55

等方面

0.54

 ingenuity

0.54

 டிர

0.54

汽車

0.53

 někol

0.53

POSITIVE LOGITS

 variant

0.96

 Production

0.91

後期

0.89

 variants

0.87

 production

0.83

型

0.81

Production

0.80

 Produktions

0.80

 early

0.80

后期

0.79

Activations Density 0.253%