Unverified50% confidenceBenchmarkExact time
On the MATH dataset, GraphEvoResearcher improves by 24 percentage points over single-pass generation and by 12 percentage points over flat reflection.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/20/2026
First Seen
Valid until: 1/18/2027
Sources
Related Entities
Related Claims
Unverified在多跳推理任务上 GraphRAG 的准确率比传统 RAG 高出 40% 以上，但索引构建时间增加 3-5 倍77% similarUnverifiedGraphSpace outperforms ordinary periodic graphs by 19.3 points overall, but leads by only 1.96 points when ordinary rendering already contains all required information, and by 46.7 points when ordinary rendering completely omits the relevant fields.72% similarUnverifiedCompared with the approach of "fine-tuning every sub-step", AdaRSI achieved a slightly higher success rate while consuming only 28% of the GPU time and 22% of the self-collected demonstration data.70% similarUnverifiedOn the same synthetic dataset, the strongest tabular baseline that removes relations and keeps the same signals achieved 81.6% accuracy, with the graph structure bringing about a 14-percentage-point improvement.69% similarUnverifiedThe statistical scope of the 21% efficiency improvement figure is not clearly specified and may be a composite score of code output volume, feature delivery speed, or engineers' subjective assessments.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1073970API
curl https://kongchang.com/api/v1/knowledge/claims/1073970MCP
get_claim(id=1073970)