INDEX

Explanations

toast and proposed

The neuron responds to language around toasting or raising a glass (e.g., “raise,” “toast,” “salute,” “Cheers”).

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

 gens

-0.79

Strut

-0.73

 Stavanger

-0.71

 Managua

-0.71

मु

-0.70

Parties

-0.70

 pesta

-0.69

 Cordoba

-0.69

fai

-0.69

 Treatise

-0.68

POSITIVE LOGITS

 toast

2.13

Toast

1.80

 Toast

1.67

toast

1.39

 toasted

1.32

showToast

1.16

 Proposed

1.09

 propose

1.09

 proposed

1.03

Proposed

1.02

Activations Density 0.012%