INDEX

Explanations

bidirectional relations, communications, and studies

This neuron detects occurrences of the term “directional,” especially in compound words like “bidirectional” or “unidirectional.”

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

棡

-2.91

 qualifié

-2.86

鷙

-2.67

⑉

-2.53

 decidedly

-2.53

dà

-2.53

 lamented

-2.47

หรือ

-2.45

 ___________

-2.45

𐄁

-2.41

POSITIVE LOGITS

and

2.81

ஂ

2.52

2.36

 After

2.31

)。

2.27

There

2.25

)，

2.22

－－

2.20

غرب

2.19

after

2.17

Activations Density 0.005%