INDEX

Explanations

have or had

The neuron is keyed to single-quote marks (apostrophes and single-quoted delimeters) in text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

لے

-0.77

 würden

-0.67

 fatiga

-0.66

ﾗ

-0.66

 باشند

-0.65

 contains

-0.65

 julie

-0.63

andar

-0.63

Being

-0.63

 Nishi

-0.63

POSITIVE LOGITS

have

1.41

 have

1.30

 hasattr

1.28

 having

1.25

 Have

1.20

 HAVE

1.16

 hebben

1.11

HAVE

1.11

toHave

1.06

avoir

1.01

Activations Density 0.037%