INDEX

Explanations

the specific noun

The neuron flags discourse-structuring cue words that signal shifts in focus or state—terms like “remaining,” “current,” or “focus.”

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 mountainous

0.37

 octagon

0.34

 mountain

0.34

 squadron

0.34

 három

0.34

odore

0.33

 peninsula

0.33

 hilly

0.33

 observatory

0.33

 chessboard

0.32

POSITIVE LOGITS

ToBe

0.36

</>

0.34

 there

0.34

ALL

0.33

and

0.33

YOU

0.33

 tất

0.33

 everytime

0.32

 นะ

0.32

하지

0.32

Activations Density 0.000%