INDEX

Explanations

expected and revealed outcomes

The neuron strongly responds to words meaning “reveal” or “disclose,” e.g. “reveal,” “revealed,” etc.

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

忝

-1.52

 perched

-1.52

 regata

-1.47

 ながら

-1.46

 tomba

-1.45

 cuillère

-1.45

 semana

-1.45

rätt

-1.42

).}

-1.41

días

-1.41

POSITIVE LOGITS

1.77

–

1.55

trä

1.54

1.49

 Два

1.45

be

1.43

 jajaran

1.39

1.38

 Twenty

1.38

此时

1.37

Activations Density 0.029%