INDEX

Explanations

`local`

The neuron fires on typical shell‐script command invocations (e.g. lines beginning with echo, grep, tail, basename, BLOCK/PART markers, etc.).

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

どれ

0.59

 wich

0.59

その他

0.58

 bata

0.56

 menentukan

0.55

 остальные

0.55

 tangente

0.55

២

0.54

 determinar

0.54

 victor

0.54

POSITIVE LOGITS

gems

0.68

dyž

0.66

Voici

0.66

 Многие

0.63

热闹

0.63

ন্‌

0.63

opération

0.62

jr

0.60

 communauté

0.59

 राज्यात

0.59

Activations Density 0.680%