INDEX

Explanations

has followed by other words

The neuron strongly activates on instances of the token “has” (including as part of identifiers like hasOwnProperty).

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

ſſel

-2.66

ſam

-2.59

ᨅ

-2.48

-2.31

 vorigen

-2.31

𝐮

-2.30

菪

-2.30

linawan

-2.30

 moreno

-2.27

 inimigos

-2.25

POSITIVE LOGITS

 been

3.28



2.58

 Could

2.45

2.39

2.34

 Focusing

2.31

 Maintaining

2.25

 Whether

2.14

 Choosing

2.14

AND

2.13

Activations Density 0.016%