INDEX

Explanations

centuries and dates

The neuron primarily activates on numeric date references (years, centuries, decades) in the text.

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

 지난

0.93

ในปี

0.88

今年的

0.86

 בשנת

0.82

since

0.78

今年

0.75

 ১৯৬৫

0.74

 છેલ્લા

0.74

 ১৯৮

0.73

早在

0.73

POSITIVE LOGITS

 onwards

1.42

 silam

1.23

 onward

1.09

0.92

′

0.91

″

0.90

′,

0.89

0.80

 itib

0.79

0.77

Activations Density 0.164%