Unverified50% confidenceFactExact time
Repetition degeneration is a common failure in large-model text generation, where the model repeatedly outputs identical or highly similar sentences.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
DeepSeek-V4.1 Meets AI Galgame: A Hands-On Test of Free Web-Based Romance
bilibili雷霆大思考10/18/2026
Related Entities
Related Claims
Unverified大语言模型在生成查找目标文本时依赖记忆重建而非逐字符复制，在长上下文场景下会出现记忆漂移导致字符串替换静默失败，此现象被称为幻觉性精确71% similarUnverifiedEmbedding models still generate vectors for grammatically broken text that are close to those for normal contract sentences, because they essentially statistically model character co-occurrence patterns rather than checking grammar.71% similarUnverifiedIn extractive summarization, every output sentence is copied directly from the input, eliminating generation-induced factual errors at the architectural level.67% similarUnverifiedMLX 不同版本间可能调整算子实现、累加顺序或低精度格式舍入策略，可能导致大模型推理输出文本的措辞变化、重复或乱码。66% similarUnverified仅仅改变输入文本的措辞和结构，就能让同一个模型的性能产生数十个百分点的差异66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1059768API
curl https://kongchang.com/api/v1/knowledge/claims/1059768MCP
get_claim(id=1059768)