INDEX

Explanations

let alone much less

The neuron detects words (especially “let” and its partner “alone,” as well as “much less”) used in negative-contrast constructions like “not X, let alone Y.”

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

 like

-1.66

we

-1.58

-1.48

 було

-1.43

-1.31

are

-1.30

同様

-1.29

asserole

-1.28

 will

-1.27

 same

-1.27

POSITIVE LOGITS

 üç

1.67

這

1.63

gni

1.52

dáms

1.46

ggy

1.46

wast

1.45

Kapcsolódó

1.44

 pasará

1.43

</b>

1.42

lya

1.41

Activations Density 0.008%