Unverified50% confidenceFactExact time
In the training corpus, bold and italics are mostly used for headings, emphasis, or typesetting rather than linguistically meaningful focus marking, which systematically shifts the model's generation distribution.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe output of a generative model is essentially a series of classification choices over a vast vocabulary, and the boundary between generation and classification is more blurred than product marketing would suggest.73% similarUnverifiedWhen the framework's default template is inconsistent with the format used during model training, model behavior drifts, including self-referential terms, refusal thresholds, response verbosity, and the way multi-turn conversational context is understood.73% similarUnverifiedCertain specific phrasings in the original prompt activate shortcuts entrenched during model training, causing the model to apply fixed patterns in its responses; after paraphrasing, the model is forced to perform more general reasoning, thereby improving output quality.72% similarUnverified模型对 schema 的遵循能力受训练数据分布影响，训练语料中高频出现的字段组合会导致模型产生幻觉字段72% similarUnverifiedThe root cause of anisotropy lies in the training objectives and architectural design of pre-trained language models (such as the BERT series), which tend to concentrate high-frequency words and general semantics in certain specific directions.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1115556API
curl https://kongchang.com/api/v1/knowledge/claims/1115556MCP
get_claim(id=1115556)