INDEX

Explanations

declaration, definition, or setting

The neuron spots code-style identifiers and macro/function names (e.g. programming or LaTeX command tokens) within 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

-3.11

-2.61

 如果

-2.58

單

-2.56

Privatpatient

-2.48

𖣘

-2.42

entikan

-2.38

熒

-2.38

雙

-2.34

cakup

-2.34

POSITIVE LOGITS

is

3.34

the

3.14

2.72

“

2.39

”]

2.39

”

2.38

es

2.36

)”

2.33

</sup>

2.28

newcommand

2.23

Activations Density 0.001%