INDEX

Explanations

negation of findings or actions

The neuron strongly activates on occurrences of the negation “not” (often in phrases like “not detected,” “not found,” or “not limited”).

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

⟣

-1.72

ęki

-1.66

྄

-1.58

跺

-1.57

 saiu

-1.56

 quitte

-1.55

 corrom

-1.52

 nuvem

-1.51

 ancha

-1.49

 emporter

-1.48

POSITIVE LOGITS

and

1.74

There

1.58

If

1.48

 赤ちゃん

1.48

 朝食

1.45

What

1.45

or

1.45

Why

1.44

錚

1.42

诸多

1.41

Activations Density 0.152%