INDEX

Explanations

subtitles, captions, dubbed audio

The neuron fires on words referring to subtitles or captioning (e.g. “subtitle(s),” “sub‐tit‐les,” “captioning,” “captions”).

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

 Aucun

-1.61

вери

-1.51

-1.48

 sinnvoll

-1.42

みて

-1.42

朦

-1.41

 brauchst

-1.41

 Aucune

-1.38

しても

-1.37

 testé

-1.36

POSITIVE LOGITS

怎么回事

1.67

 nasz

1.59

esticular

1.52

靦

1.48

 importance

1.45

?”,

1.44

.},

1.44

</h1>

1.42

 ΤΟ

1.41

かもしれません

1.40

Activations Density 0.007%