INDEX

Explanations

all [noun]

The neuron strongly activates on the universal quantifier “all.”

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

いた

1.52

ם

1.41

cı

1.40

ία

1.31

rifice

1.31

daki

1.30

ように

1.30

 awfully

1.28

ка

1.27

сть

1.27

POSITIVE LOGITS

 kecuali

1.38

usions

1.34

igators

1.33

iteration

1.29

 demás

1.24

 تقريبا

1.24

 kinds

1.23

ergic

1.18

 sorts

1.17

uding

1.09

Activations Density 0.328%