INDEX

Explanations

product options and listings

The neuron lights up on programming‐style identifier tokens (e.g. function names, class or variable names, and other code keywords).

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

Indirect

-0.73

 Cust

-0.72

 Bounce

-0.70

jiga

-0.66

籽

-0.66

 accelerators

-0.66

 событий

-0.65

Van

-0.65

 Lenn

-0.65

 inspire

-0.64

POSITIVE LOGITS

niaus

0.94

刊

0.84

 phí

0.84

 />);

0.83

psi

0.77

 specifics

0.77

 Catedral

0.74

Tpl

0.74

гъл

0.73

 Nantes

0.73

Activations Density 0.001%