INDEX

Explanations

The neuron strongly activates on code “for … loop” constructs—i.e. occurrences of a for-loop keyword.

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

 connaissances

0.60

लीज

0.59

ின்

0.58

太过

0.57

成长

0.56

 निर्

0.55

malıdır

0.55

是什么

0.54

্বল

0.54

最新

0.53

POSITIVE LOGITS

idx

1.49

1.44

 index

1.41

 índice

1.41

index

1.29

 indices

1.29

idx

1.27

Idx

1.26

 indeks

1.26

 indexes

1.23

Activations Density 0.031%