INDEX

Explanations

hugging face, urgency, guarantee

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

 hijacking

2.89

 coolness

2.79

monium

2.67

таки

2.64

 wavelet

2.61

xty

2.57

 ocorre

2.55

ons

2.52

駁

2.44

 cool

2.44

POSITIVE LOGITS

<bos>

3.96

prising

3.19

propion

3.05

ኔታ

2.97

র্বর

2.94

πηρε

2.93

uu

2.92

ष्मान

2.82

клад

2.80

ोहित

2.77

Activations Density 2.623%