INDEX

Explanations

indicating origin

The neuron specifically fires on the word “derived,” as used in compound terms (e.g. “stem‐cell–derived,” “cell culture–derived,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

</h1>

-3.73

</i>

-3.30

was

-3.00

ти

-2.88

or

-2.86

꜆

-2.80

 стал

-2.75

鑁

-2.38

?"

-2.36

."

-2.28

POSITIVE LOGITS

𐄁

3.22

፨

2.67



2.64

浹

2.56

他

2.52

 realizará

2.50

 spolup

2.47

 presentará

2.45

 összefoglaló

2.41

Продам

2.30

Activations Density 0.008%