INDEX

Explanations

you will need

The neuron selectively activates on common function words—especially pronouns (I, we), auxiliaries and modals (would, is, to)—rather than content words.

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

 हैं

0.89

 आहेत

0.86

 Están

0.83

 है

0.82

 आहे

0.79

 estão

0.77

 Είναι

0.77

 Está

0.76

 έχουν

0.75

 êtes

0.73

POSITIVE LOGITS

也要

0.72

やはり

0.69

 другой

0.64

 другую

0.62

 procurar

0.61

 therapeutics

0.59

 вопрос

0.58

 desirous

0.58

 innervation

0.58

 জিজ্ঞাসা

0.57

Activations Density 0.459%