INDEX

Explanations

The neuron fires on filesystem path tokens (e.g. directory names, slashes, device paths).

configuration and doing

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

 muitos

-0.89

 molte

-0.83

 many

-0.83

₅

-0.82

 leftover

-0.81

ateľ

-0.81

 wichtigste

-0.79

Abbiamo

-0.79

 kathol

-0.79

 muchas

-0.79

POSITIVE LOGITS

thebibliography

0.97

utilisons

0.91

 文章

0.90

windig

0.89

vasive

0.89

 apps

0.88

ﺀ

0.88

Addresses

0.88

Спорт

0.86

😕

0.86

Activations Density 0.331%