INDEX

Explanations

communicate or help

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

kunjung

-0.94

)}]{

-0.93

грі

-0.82

 jurk

-0.80

ORCID

-0.79

 homolog

-0.79

Meals

-0.79

 Normalize

-0.78

hadapi

-0.77

Madam

-0.77

POSITIVE LOGITS

 help

1.23

 communicated

1.22

 communicate

1.16

 communication

1.15

 communicating

1.08

 communicates

1.06

help

1.01

Help

1.00

 position

0.98

 positions

0.96

Activations Density 0.006%