林煒清 | Wei-Ching Lin
weiching-lin.bsky.social
林煒清 | Wei-Ching Lin
@weiching-lin.bsky.social
data, data, and data
i think it is still useful for summarization type of question?

i saw it perform badly here. the Figure 1 shows GraphRAG underperform large embedding model on all types of question.
www.alphaxiv.org/abs/2502.14802
From RAG to Memory: Non-Parametric Continual Learning for Large Language Models | alphaXiv
View 5 comments: It is good to know how much questions that HippoRAG is turned into simple RAG due to Triple Filtering.So we can know how important the component is.I guess it would be the key compone...
www.alphaxiv.org
June 27, 2025 at 10:26 AM