INDEX

Explanations

of followed by article or possessive

The neuron activates on numeric expressions (counts, measures, coefficients) wherever they appear in the 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

 voire

-1.34

 barca

-1.32

roj

-1.31

.",

-1.26

rua

-1.24

-1.23

 luxo

-1.22

 SOMETHING

-1.21

 comodo

-1.21

()]

-1.20

POSITIVE LOGITS

 recomendamos

1.52

いいのか

1.52

 όταν

1.41

 competitivo

1.39

 quando

1.38

 esperan

1.34

pēc

1.31

 essas

1.28

tiqueta

1.26

 когда

1.24

Activations Density 0.066%