INDEX

Explanations

end of block

The neuron specifically fires on TeX “\end” tokens (i.e. the ends of environments or the document).

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

鹋

-1.99

to

-1.80

老子

-1.66

 yine

-1.61

 this

-1.59

harmed

-1.59

那你

-1.59

 myös

-1.56

 menino

-1.56

等は

-1.54

POSITIVE LOGITS

↵

1.89

end

1.66

1.63

 estekak

1.60

 tredje

1.58

。

1.57

}}$

1.50

}$.

1.49

plantilla

1.49

1.48

Activations Density 0.001%