Unverified50% confidenceFactExact time
paraphrase-multilingual-MiniLM-L12-v2 reads at most 128 tokens, and the excess is silently truncated, which is one of the reasons RAG needs to split documents into chunks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/13/2026
First Seen
Valid until: 1/11/2027
Sources
Related Entities
Related Claims
UnverifiedFor a published 406M-parameter Fusion-in-Decoder specialized model, the ROUGE-1 score dropped sharply by 6.30 points when the input was switched from truncated long text to 2,000-word retrieved segments.65% similarUnverifiedparaphrase-multilingual-MiniLM-L12-v2 is a multilingual sentence embedding model provided by the Sentence Transformers library; it compresses the output of a 12-layer Transformer encoder into a 384-dimensional vector through mean pooling and supports about 50 languages.62% similarUnverified早期Stable Diffusion系列以CLIP为文本编码器，对超过77个Token的提示词处理能力受限62% similarUnverifiedText chunking granularity that is too large results in retrieved text containing large amounts of irrelevant information, while granularity that is too small may lose contextual semantics.60% similarUnverifiedCLIP文本编码器的标准上下文窗口长度约为77个Token，超出部分的语义信息会被截断或降权处理59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1014334API
curl https://kongchang.com/api/v1/knowledge/claims/1014334MCP
get_claim(id=1014334)