INDEX

Explanations

giving and receiving feedback

The neuron activates on occurrences of the word “feedback.”

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

-3.19

-2.75

 необходимо

-2.64

 eine

-2.63

↵↵

-2.59

 где

-2.56

 WITH

-2.53

-2.45

 beantworten

-2.42

 avec

-2.34

POSITIVE LOGITS

 parachoque

3.36

赑

3.16

狆

2.98

虓

2.88

取决

2.84

鞴

2.80

鸱

2.80

ではでは

2.78

凩

2.72

warnai

2.70

Activations Density 0.004%