{"slug": "tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting", "title": "Tracing where LLM \"slop\" style comes from in the model and in training, and cutting it out [R]", "summary": "A researcher traced the recognizable LLM assistant writing style — em dashes, \"it's not X, it's Y\" constructions, lists of three and buzzwords — to a difference-in-means direction in Qwen3-4B-Instruct, computed from the model's own exaggerated versus plain writing. The work also set out to identify which training stage installs the style and to remove it from the model's output.", "body_md": "I wanted to find where the recognizable assistant style (em dashes, \"it's not X, it's Y\", lists of three, buzzwords) lives in a model and which training stage installs it. Where it lives: in Qwen3-4B-Instruct, a difference-in-means direction from the model's own exaggerated vs plain writing separate", "url": "https://wpnews.pro/news/tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting", "canonical_source": "https://aiflash.com/news/134471/", "published_at": "2026-10-17 21:00:03+00:00", "updated_at": "2026-10-17 21:07:12.876100+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "machine-learning", "artificial-intelligence"], "entities": ["Qwen3-4B-Instruct"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting", "markdown": "https://wpnews.pro/news/tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting.md", "text": "https://wpnews.pro/news/tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting.txt", "jsonld": "https://wpnews.pro/news/tracing-where-llm-slop-style-comes-from-in-the-model-and-in-training-and-cutting.jsonld"}}