INDEX

Explanations

words starting with De

This neuron detects the token “De” (an uppercase “D” followed by “e”) appearing as a prefix in words.

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

-3.48

</h3>

-3.03

to

-2.98

zembro

-2.89

<bos>

-2.88

the

-2.83

.,"

-2.78

-2.77

eds

-2.73

."),

-2.70

POSITIVE LOGITS

3.16



2.83

芣

2.61

唻

2.59

鲻

2.52

ーーーー

2.48

 einige

2.48

 ſont

2.42

 étrange

2.41

Français

2.38

Activations Density 0.007%