INDEX

Explanations

someone else's

The neuron fires on words that denote specific other people or individuals (e.g. “Mom,” “neighbor,” “someone,” “people”)—i.e. references to particular persons.

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

 والذي

0.49

 répét

0.46

வில்

0.44

 Fargo

0.44

 آن

0.43

 primo

0.43

точ

0.42

 ছেলেকে

0.42

 নিজে

0.42

 того

0.41

POSITIVE LOGITS

들의

0.84

们的

0.74

たちの

0.67

ның

0.62

ങ്ങളുടെ

0.61

ിയുടെ

0.59

ımızın

0.58

분의

0.58

らの

0.57

ගේ

0.57

Activations Density 0.000%