INDEX

Explanations

emotion

The neuron is specialized in detecting the word “emotion” (and its close variants such as “Emotions,” “emotional,” etc.) in the text.

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

With

-2.84

Only

-2.75

as

-2.59

 esteemed

-2.58

This

-2.50

鐿

-2.50

 Honestly

-2.44

</h3>

-2.41

While

-2.39

未完

-2.38

POSITIVE LOGITS

3.72

3.67

3.22

↵

2.86

みそ

2.36

↵↵

2.22

 acresc

2.20

 funcionam

2.17

肜

2.16

The

2.14

Activations Density 0.002%