INDEX

Explanations

import, codes, title, us

The neuron fires on numeric tokens—i.e. literal numbers or version‐style digits—in code.

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

 प्रोड

-1.48

 eccl

-1.45

 even

-1.43

研究生

-1.30

 religieux

-1.30

 although

-1.30

 tucson

-1.26

ׇ

-1.25

 τὸν

-1.25

 serra

-1.25

POSITIVE LOGITS

 бампер

2.17

ίη

1.59

 Meningkatkan

1.57

 limpiar

1.51

͏

1.48

וּ

1.45

ContentAlignment

1.45

 reposer

1.41

 trám

1.41

lê

1.41

Activations Density 0.010%