INDEX

Explanations

describing context or clarification

This neuron fires on qualifier adjectives like “simple” or “typical” (and similar words such as “bright”) that introduce or characterize explanatory phrases.

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

of

-2.61

-2.20

which

-2.00

这个

-1.91

-1.87

-1.76

-1.74

for

-1.63

in

-1.62

POSITIVE LOGITS

 accordance

1.90

鶿

1.85

and

1.84

 cantidades

1.77

 alberi

1.69

 persino

1.67

嫋

1.64

 også

1.63

 tambi

1.63

 发型

1.62

Activations Density 0.156%