INDEX

Explanations

satisfy conditions and rules

The neuron fires on numerical expressions—digits, numbers (including decimals), and equation-like numeric tokens.

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

姈

-1.64

 некоторые

-1.58

be

-1.48

觊

-1.48

ACTED

-1.47

ostí

-1.45

嬲

-1.44

鸲

-1.41

掳

-1.39

 čty

-1.38

POSITIVE LOGITS

 militare

1.71

1.57

 myös

1.50

 kollu

1.50

理由は

1.43

 рослин

1.40

 universale

1.39

 gives

1.38

بدون

1.38

 malang

1.35

Activations Density 0.035%