INDEX

Explanations

units of measurement and time

The neuron consistently lights up on numeric tokens (multi‐digit numbers or measurements) 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

to

-1.48

が

-1.43

みてください

-1.38

龊

-1.35

-1.34

同样

-1.23

pett

-1.23

 extiende

-1.23

 sogen

-1.20

 terceiro

-1.17

POSITIVE LOGITS

 norvég

1.67

فران

1.45

贽

1.32

 réjou

1.28

 biela

1.27

 οποία

1.27

 while

1.27

 amélior

1.26

 espé

1.25

řízen

1.25

Activations Density 0.084%