INDEX

Explanations

Updated with date/time

The neuron detects article metadata markers—especially “Updated:” labels, dates, times, and their associated punctuation.

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

you

-1.50

他又

-1.45

不断

-1.38



-1.36

言えば

-1.36

他就

-1.36

了几

-1.34

当时的

-1.33

-1.30

居る

-1.30

POSITIVE LOGITS

<>());

1.65

-------------</

1.59

儛

1.52

斯的

1.49

britannien

1.48

<",

1.48

,‎

1.45

夀

1.43

܇

1.42

潆

1.42

Activations Density 0.027%