INDEX

Explanations

X and Y

The neuron lights up on content‐bearing, often multi‐syllabic topic words (key nouns and verbs) rather than on function words.

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



-1.61

 even

-1.59

];

-1.38

 даже

-1.38

].

-1.28

it

-1.28

 také

-1.28

₱

-1.27

其实

-1.27

 actually

-1.26

POSITIVE LOGITS

and

1.95

 zonder

1.52

but

1.52

 znalaz

1.45

 після

1.44

 sebelum

1.39

économ

1.39

piram

1.38

mushroom

1.38

venez

1.38

Activations Density 0.123%