Em Dash Overuse Pattern
(Redirected from Long Dash Pattern)
Jump to navigation
Jump to search
A Em Dash Overuse Pattern is a llm writing marker that exhibits excessive em dash usage in ai-generated text.
- AKA: Em Dash Spam, Dash Overuse, Long Dash Pattern, — Overuse, Em Dash Tell.
- Context:
- It can typically manifest at 2-3x higher frequency than human writing baselines.
- It can typically appear as parenthetical insertions and clause separators.
- It can often indicate gpt model output particularly from instruction-tuned models.
- It can often correlate with formal writing style in ai-generated content.
- It can range from being a Subtle Em Dash Overuse Pattern to being an Obvious Em Dash Overuse Pattern, depending on its frequency deviation.
- It can range from being a Consistent Em Dash Overuse Pattern to being a Variable Em Dash Overuse Pattern, depending on its occurrence regularity.
- It can range from being a Model-Specific Em Dash Overuse Pattern to being a Universal Em Dash Overuse Pattern, depending on its llm scope.
- It can range from being a Context-Dependent Em Dash Overuse Pattern to being a Context-Independent Em Dash Overuse Pattern, depending on its domain sensitivity.
- ...
- Examples:
- GPT-4 Em Dash Patterns, such as:
- "The results—which were surprising—indicate that..." occurring with high frequency.
- "The system—a complex architecture—demonstrates..." as characteristic construction.
- Claude Em Dash Patterns, such as:
- Multiple em dashes per paragraph for thought separation.
- Em dash clusters in explanatory text.
- Academic Writing Em Dash Patterns showing formal tone markers.
- ...
- GPT-4 Em Dash Patterns, such as:
- Counter-Examples:
- Hyphen Usage Pattern, which uses short dashes for word connection.
- En Dash Usage Pattern, which uses medium dashes for range indication.
- Parenthesis Usage Pattern, which uses bracket punctuation rather than dash punctuation.
- Natural Em Dash Usage, which follows human writing conventions.
- See: LLM Writing Marker, Punctuation Pattern, AI-Generated Text Detection Task, Stylometric Analysis, GPT Model Output Pattern, Claude Model Output Pattern, Writing Style Analysis, Syntactic Pattern, Text Authenticity Measure.