INDEX

Explanations

failure, mistakes, imperfection

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

Items

0.87

elemento

0.84

items

0.82

interactions

0.79

ồi

0.77

issue

0.77

마트

0.76

Util

0.76

engkapi

0.75

stap

0.75

POSITIVE LOGITS

 damned

1.32

 damn

1.19

 fucking

1.15

ain

1.14

 proverbial

1.14

 ugly

1.13

 roses

1.12

 awfully

1.09

 mighty

1.07

 darn

1.06

Activations Density 2.696%