INDEX

Explanations

research papers and institutions

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

ibu

1.56

éraux

1.52

rul

1.51

낀

1.51

rens

1.50

庆祝

1.50

 heureux

1.48

tras

1.47

éristiques

1.47

rectomy

1.46

POSITIVE LOGITS

ipient

2.19

ington

2.18

成果

2.13

 ActionResult

2.05

يّ

2.04

 flasks

2.03

Purple

1.99

aust

1.99

itej

1.98

真っ

1.97

Activations Density 0.055%