INDEX

Explanations

npm install, yarn add

The neuron strongly activates on numeric tokens (e.g. standalone numbers, digits, and numeric patterns).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 দেয়নি

0.56

因

0.54

 Aank

0.54

违

0.52

郫

0.52

 palav

0.52

োঁ

0.52

 Frag

0.52

Faj

0.52

क्षकों

0.52

POSITIVE LOGITS

Private

0.68

private

0.66

license

0.65

readme

0.64

ENGINE

0.64

magnetic

0.63

engines

0.62

License

0.60

PRIVATE

0.60

Engine

0.60

Activations Density 0.001%