INDEX

Explanations

code block beginnings

The neuron primarily detects filenames or resource names with programming‐related file extensions (e.g. “.html,” “.sh,” “.php”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 throwIf

0.63

рактеристики

0.63

щают

0.59

 sortBy

0.58

 adversaries

0.55

 Plates

0.55

 utils

0.54

istet

0.54

 KEYS

0.54

 கள

0.51

POSITIVE LOGITS

<html>

1.28

documentclass

1.01

DOCTYPE

0.89

<!

0.87

#!/

0.76

<!

0.76

<?

0.68

DOCUMENT

0.67

Document

0.62

UTF

0.61

Activations Density 0.019%