INDEX

Explanations

concept

The main thing this neuron does is detect occurrences of the word “concept” (including its plural form and high-activation variants) 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

the

-2.61

 フラワー

-2.50

栳

-2.42

趿

-2.42

."

-2.42

焢

-2.38

鬭

-2.34

 pengusaha

-2.20

</i>

-2.17

糌

-2.17

POSITIVE LOGITS

”،

3.52

</em>

3.11

3.05

</h2>

3.00

</sup>

2.94

</h4>

2.83

’

2.72

et

2.70

</h6>

2.69

<em>

2.56

Activations Density 0.018%