INDEX

Explanations

word fragments then completions

The neuron strongly activates on multi‐digit numeric tokens (e.g. years or long strings of digits).

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

 such

-1.57

you

-1.50

so

-1.42

 several

-1.40

it

-1.40

for

-1.37

 provide

-1.37

 many

-1.32

on

-1.30

 this

-1.30

POSITIVE LOGITS

ὶ

1.44

 reflexiones

1.44

┍

1.44

𝑄

1.43

Vastaa

1.37

 рецепта

1.35

 schuhe

1.31

Torna

1.29

 bivolt

1.27

ederen

1.25

Activations Density 0.000%