INDEX

Explanations

items, grades, nationalities

The neuron fires most strongly on four-digit year tokens (i.e. it detects year numbers).

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

 này

0.60

 onboarding

0.54

 reactivity

0.52

ষে

0.51

 imperme

0.51

 moisture

0.50

这个

0.49

 initializing

0.49

 accuracy

0.49

 this

0.48

POSITIVE LOGITS

 Diarsipkan

0.70

</td>

0.63

あるいは

0.58

 혹은

0.57

 Biographie

0.56

もしくは

0.55

 naciones

0.54

 Перейти

0.54

 Internacional

0.52

 Oyun

0.52

Activations Density 0.000%