INDEX

Explanations

proposed or adopted solutions

This neuron detects mentions of proposed technical approaches or implementations—terms like “solutions,” “method,” “system,” “scheme,” or “design” in patent-style descriptions.

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

岌

-1.67

</strong>

-1.34

-1.33

都可以

-1.30

mathrm

-1.30

 světa

-1.26

Begriff

-1.25



-1.19

 ezen

-1.16



-1.16

POSITIVE LOGITS

 选择

1.44

 DETAILS

1.38

 scelta

1.33

を採用

1.30

example

1.27

depan

1.24

 bumbu

1.23

 通过

1.23

たり

1.21

ウザ

1.21

Activations Density 0.084%