INDEX

Explanations

persuading or convincing

This neuron detects tokens from verbs expressing persuasion or convincing (e.g., forms of “persuade,” “convince,” “talk someone into,” “hire”).

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

any

-1.47

Essex

-1.39

 cualquier

-1.33

 Schö

-1.31

 they

-1.30

 which

-1.27

 whose

-1.25

 которых

-1.21

 still

-1.20

VENTION

-1.19

POSITIVE LOGITS

 poética

1.38

 några

1.37

 вполне

1.34

 akibat

1.32

 halben

1.31

 вывод

1.30

𝟑

1.29

 respective

1.27

 гораздо

1.27

 Polícia

1.25

Activations Density 0.061%