INDEX

Explanations

route parameters

The neuron is detecting tokens composed of numeric digit sequences—e.g. the numeric IDs or parameter placeholders in URL routes.

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

ledad

-1.44

などを

-1.40

 technisch

-1.34

umers

-1.31

や

-1.30

EDTA

-1.27

を

-1.27

 remarquer

-1.23

 manteiga

-1.23

และ

-1.20

POSITIVE LOGITS

で見る

1.67

場所で

1.66

にしました

1.36

 overtly

1.33

ほら

1.33

にします

1.28

に挑戦

1.26

僱

1.25

為に

1.25

One

1.25

Activations Density 0.002%