INDEX

Explanations

image and screenshot references

The neuron activates on image references and figure placeholders (e.g., “screenshot,” bracketed figure numbers, “PICTURE,” or Markdown/URL image links).

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

 οποίο

-1.05

⌒

-0.97

낱

-0.91

 Detailed

-0.91

opje

-0.90

bertura

-0.89

||}

-0.89

rý

-0.88

させていただきます

-0.87

 Bearbeitung

-0.86

POSITIVE LOGITS

 images

1.15

图

1.09

 già

1.05

 where

0.97

在这里

0.96

 imagenes

0.92

эль

0.91

Figure

0.90

image

0.90

 cliquez

0.89

Activations Density 0.051%