INDEX

Explanations

signatures and authors

The neuron is selectively activating on numeric tokens—especially decimal‐point numbers—throughout the document.

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

but

-1.52

mathrm

-1.48

 beschäf

-1.43

 this

-1.36

 ухода

-1.30

 oppure

-1.29

 that

-1.27

ותו

-1.25

 αλλά

-1.24

せい

-1.23

POSITIVE LOGITS

蒾

1.48

 каждой

1.41

 have

1.40

说着

1.40

は

1.38

位于

1.37

 constaté

1.37

푉

1.35

从

1.34

By

1.34

Activations Density 0.084%