INDEX

Explanations

The neuron is sensitive to content‐bearing words (i.e. non–stop words such as nouns, descriptive verbs, numbers, and proper names) rather than function words.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

さまざまな

0.93

Kaynakça

0.82

を有する

0.82

当該

0.81

Alternatively

0.77

Asimismo

0.75

 нередко

0.75

Alongside

0.73

निम्नलिखित

0.72

Capability

0.72

POSITIVE LOGITS

 soooo

1.85

 pero

1.80

cuz

1.73

 sooo

1.73

!!!!

1.71

but

1.65

!!

1.61

!!!

1.60

!!!!!

1.59

 haha

1.58

Activations Density 6.210%