INDEX

Explanations

for a purpose

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

筚

1.84

 shackles

1.76

oules

1.70

ovati

1.66

actives

1.63

 takers

1.61

続く

1.58

どうぞ

1.58

ător

1.57

lepiej

1.56

POSITIVE LOGITS

某

2.24

several

2.03

 another

2.00

 beberapa

1.95

another

1.93

 friend

1.91

apparently

1.87

 saya

1.84

one

1.84

 colleague

1.82

Activations Density 0.210%