INDEX

Explanations

award of the year

The neuron sharply activates on high‐frequency function words—especially the definite article “the” (and to a lesser extent “of”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 passende

-1.63

 lidé

-1.59

完成です

-1.57

杙

-1.57

 rozdí

-1.55

 dobré

-1.52

が無い

-1.46

 adaptés

-1.45

 standen

-1.44

POSITIVE LOGITS

Those

1.66

Also

1.59

and

1.59

 However

1.49

However

1.46

Additionally

1.46

There

1.45

 Therefore

1.44

そちら

1.41

/>

1.41

Activations Density 0.015%