INDEX

Explanations

sliders / s / SL

The neuron fires on tokens that look like alphanumeric identifiers or acronyms—i.e. sequences of uppercase letters (often mixed with digits) such as model numbers, codes, or shorthand IDs.

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

-1.38

 extensive

-1.36

-1.32

cuatro

-1.31

You

-1.26

 lustig

-1.26

-1.25

</strong>

-1.24

 Bruce

-1.24

ten

-1.21

POSITIVE LOGITS

;&#

1.52

🧌

1.30

nachron

1.30

 Ministre

1.28

のほうが

1.26

 chaussures

1.24

 ALUM

1.24

 PUST

1.23

 cardin

1.22

 Tind

1.22

Activations Density 0.166%