INDEX

Explanations

starts with date, a, recently, or Q

This neuron detects opening delimiters in markup or code (like “{” in math environments or “<” in HTML tags).

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

</b>

-2.33

 ursprünglichen

-2.02

 also

-1.99

尤其

-1.95

</h4>

-1.95

 или

-1.89

Чтобы

-1.87

-1.81

 übrigen

-1.78

POSITIVE LOGITS

锨

2.17

2.09

γραφή

2.05

ਞ

2.05

)});

1.96

ikuti

1.86

丏

1.84

鲠

1.82

 occafion

1.81

hafter

1.79

Activations Density 0.000%