INDEX

Explanations

multilingual or technical contexts

The neuron is keyed to the “#” symbol, effectively picking up on hash‐prefixed headings or comment markers.

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

 неиз

0.84

Campe

0.84

Conj

0.82

Malformed

0.81

ν

0.80

0.79

Comic

0.77

Slf

0.77

 día

0.77

POSITIVE LOGITS

 subscrib

0.80

पृथ

0.80

కీయ

0.79

থিল

0.78

啇

0.77

 amending

0.76

 accru

0.76

᱕

0.76

sunline

0.75

 koordin

0.75

Activations Density 0.001%