INDEX

Explanations

version

This neuron selectively activates on isolated content‐bearing nouns (or noun‐like headings) that introduce or label a particular category or section.

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

 sehingga

0.71

 což

0.61

 selaku

0.58

Le

0.56

 kepada

0.54

മുണ്ട

0.52

 बदले

0.51

克兰

0.49

 Leute

0.48

 kişi

0.48

POSITIVE LOGITS

 المتوسط

0.59

 solamente

0.56

মাদের

0.55

 relembrar

0.55

AND

0.54

一般的

0.54

ongen

0.53

 frowning

0.53

 Timelapse

0.53

uncia

0.52

Activations Density 0.001%