INDEX

Explanations

React code following Q or Introduction

The neuron strongly activates on programming/code syntax elements (identifiers, punctuation, keywords, URLs) indicating it’s detecting code blocks.

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

 Leicestershire

-0.96

Sow

-0.79

 ressource

-0.77

Wab

-0.75

 Forder

-0.75

anius

-0.75

 Brahmin

-0.74

 చే

-0.73

Coh

-0.73

Bills

-0.71

POSITIVE LOGITS

 React

1.02

 className

0.90

Kde

0.86

ClassName

0.83

zve

0.81

 useCallback

0.81

pula

0.80

sumoto

0.79

 react

0.78

 весьма

0.78

Activations Density 0.015%