INDEX

Explanations

node; next; left; right

This neuron fires on pointer‐field accesses in data structures—especially the “→next” links in lists (and similarly “→left/→right” in trees).

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

 that

-1.87

 their

-1.69

 which

-1.64

 there

-1.47

 simply

-1.45

if

-1.42

 их

-1.40

you

-1.39

now

-1.33

 start

-1.28

POSITIVE LOGITS

 koste

1.39

zwei

1.37

 kredi

1.34

 konser

1.33

PLICATIONS

1.30

ᨆ

1.30

 italiens

1.30

Mujer

1.27

 lidé

1.27

 parfü

1.26

Activations Density 0.007%