INDEX

Explanations

introduces subordinate clauses

The neuron fires on subjective, evaluative words and phrases—i.e. terms that express opinion or judgment (like “super‐critical reader,” “nice person,” “just,” “interesting,” “customer,” 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

 significantly

0.65

 جائے۔

0.57

 यामध्ये

0.57

brar

0.55

 यात

0.55

 ہونے

0.54

 மட்டும்

0.53

 ovom

0.53

 acest

0.52

넌

0.52

POSITIVE LOGITS

 that

4.63

that

4.44

 bahwa

3.95

 rằng

3.69

 bahawa

3.67

 THAT

3.44

 kwamba

3.41

That

3.38

THAT

3.36

 दैट

3.27

Activations Density 1.435%