INDEX

Explanations

interaction with

The neuron detects multiword phrasal‐verb constructions—that is, a verb combined with a particle or preposition (e.g. “facing off with,” “taking on,” “going up against,” “talks with”).

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

ulipas

-1.58

Nachdem

-1.51

 színes

-1.48

 amikor

-1.48

skapet

-1.46

)$,

-1.44

čki

-1.41

ilización

-1.40

ógicas

-1.39

 női

-1.39

POSITIVE LOGITS

its

2.08

 other

2.05

 only

1.98

 their

1.83

 with

1.70

on

1.62

 between

1.56

1.52

for

1.52

 some

1.52

Activations Density 0.224%