INDEX

Explanations

searching verbs and concepts

The neuron activates strongly on uppercase configuration keywords, underscores, hyphens, and other punctuation as found in CMake‐style directives (e.g. “INCLUDE_DIRS,” “FIND_PATH,” “-”)

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

EVAL

-0.81

 наблю

-0.78

 lấy

-0.77

评

-0.76

anci

-0.76

zhen

-0.76

onSelect

-0.75

 thiệu

-0.75

vikling

-0.73

ูด

-0.73

POSITIVE LOGITS

 search

3.39

 searching

2.89

 searched

2.75

 searches

2.75

search

2.58

 Search

2.45

Search

2.44

搜索

2.33

Searching

2.23

 SEARCH

2.20

Activations Density 0.110%