INDEX

Explanations

by [agent performing action]

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

 natoque

2.88

隍

2.77

FileAlignment

2.75

ций

2.72

ಷ

2.68

ীল

2.68

𝙢

2.61

ikel

2.57

্যান্ড

2.52

زاده

2.50

POSITIVE LOGITS

ا

3.30

reise

2.93

ringstream

2.91

2.73

ه‌ی

2.69

et

2.63

সিডেন্ট

2.62

اٹ

2.59

্ম্ম

2.58

ன

2.56

Activations Density 0.131%