INDEX

Explanations

random number, speed, patch

The neuron strongly activates on numeric measurements—especially floating‐point numbers and other explicit numerical values.

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

 kaip

0.88

 Ajay

0.88

jub

0.84

<unused94>

0.82

 ligados

0.82

 mitophagy

0.76

að

0.76

 QnrB

0.76

 triglycer

0.74

 ziff

0.74

POSITIVE LOGITS

etc

0.91

label

0.88

and

0.81

live

0.76

lass

0.75

la

0.74

aircraft

0.74

he

0.73

ag

0.73

이면

0.72

Activations Density 0.000%