INDEX

Explanations

The neuron detects formal legal or administrative terminology and section‐heading phrases typical of policies or terms-of-service documents.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sellest

0.73

ತಿಯ

0.68

 afin

0.64

このように

0.63

ಿಂತ

0.61

اہم

0.59

之中

0.57

FromThe

0.57

ரத்தில்

0.57

を中心に

0.57

POSITIVE LOGITS

 для

5.31

for

5.19

สำหรับ

4.66

 für

4.63

 για

4.63

для

4.56

 untuk

4.47

for

4.38

 voor

4.31

 برای

4.28

Activations Density 4.355%