INDEX

Explanations

division and calculation

The neuron strongly activates on large numeric literals—especially multi‐digit decimal numbers (e.g. “1277.67,” “1245.4763,” etc.).

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

Asimismo

0.31

IMPLEMENT

0.27

либо

0.27

 издели

0.27

たとえば

0.26

 পরিচালিত

0.26

 말미암아

0.26

Hãy

0.25

 veroor

0.25

คาเดมี

0.25

POSITIVE LOGITS

but

0.56

 pero

0.49

 αλλά

0.46

ですが

0.45

 nhưng

0.44

但是

0.44

 কিন্তু

0.43

但是我

0.43

but

0.43

 కానీ

0.42

Activations Density 0.000%