INDEX

Explanations

what to do or know

The neuron is highly sensitive to the token “to,” especially when it appears as the infinitive marker in phrases like “what to …” or “how to ….”

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

巨大的

-1.92

<bos>

-1.80

独特的

-1.71

astify

-1.70

 είναι

-1.63

与此同时

-1.59

葑

-1.57

isation

-1.55

 ostavi

-1.55

 optimieren

-1.54

POSITIVE LOGITS

 μην

2.05

be

1.80

1.51

 know

1.45

1.43

だったり

1.38

1.37

 cuando

1.36

el

1.34

But

1.34

Activations Density 0.036%