INDEX

Explanations

PLOS ONE citations

This neuron detects journal-title tokens in reference citations (e.g. “PLoS,” “ONE,” “pone,” etc.).

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

訁

-1.81

 ส์

-1.77

䜣

-1.69

蛱

-1.68

 บ้าน

-1.67

 naszym

-1.66

 oído

-1.61

we

-1.61

 ſont

-1.59

 คา

-1.59

POSITIVE LOGITS

 while

1.88

</em>

1.81

</sup>

1.77

1.63

 more

1.60

 both

1.60

еры

1.59

 allowing

1.54

 allows

1.52

ències

1.49

Activations Density 0.014%