INDEX

Explanations

Mr. / Ms. salutations

This neuron flags metadata-like tokens such as dates, headings, or field labels (e.g. month names, "Entry," "filed under," "complete").

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

ية

0.81

不到

0.73

した後

0.73

 finest

0.70

பதி

0.70

 fiss

0.69

 znovu

0.68

গর্

0.67

センター

0.67

ত

0.67

POSITIVE LOGITS

[{{

0.84

 വേദന

0.82

icycles

0.81

 pijn

0.80

 xấu

0.80

💊

0.80

rylic

0.79

 väl

0.79

 يلا

0.77

 YEARS

0.77

Activations Density 0.001%