INDEX

Explanations

defines or states a premise

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 representan

-0.82

alyptus

-0.81

UTES

-0.81

 mewakili

-0.78

koper

-0.76

zara

-0.75

represented

-0.74

zension

-0.74

behör

-0.72

Владимир

-0.72

POSITIVE LOGITS

 states

3.58

 stating

3.48

 stated

3.16

 says

2.89

 stip

2.38

states

2.36

 saying

2.17

stated

2.03

 stipul

1.98

 STATES

1.88

Activations Density 0.116%