INDEX

Explanations

This neuron strongly fires on isolated code/math identifiers (single letters or numbers) used as function arguments or variable names in formulas.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 হ্

0.57

osa

0.57

 संसा

0.55

Myers

0.55

ຕິດຕໍ່

0.55

."},

0.55

"":

0.55

 fastened

0.55

 interacción

0.55

 परस्पर

0.54

POSITIVE LOGITS

,{

1.37

,(

1.27

,[

1.18

);

1.09

,[

1.09

，

1.09

',{

1.08

',

1.06

1.05

',['

1.02

Activations Density 0.023%