Beyond Chunks and Graphs: Retrieval-Augmented Generation through Triplet-Driven Thinking
Shengbo Gong, Xianfeng Tang, Carl Yang, Wei jin · 2025
Read Beyond Chunks and Graphs: Retrieval-Augmented Generation through Triplet-Driven Thinking by Shengbo Gong, Xianfeng Tang, Carl Yang, Wei jin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-02435.txt.
Description
Shengbo Gong and co-authors wrote this 2025 paper on a leaner way to feed facts to language models. Retrieval systems either loop many times, burning costly model calls, or build knowledge graphs that are slow to construct and error-prone. The alternative from the team skips graphs entirely and stores knowledge as atomic triplets, simple subject-predicate-object facts.
Their framework, T-squared RAG, has the model rewrite each question as triplets with blanks, then fills the blanks step by step from the triplet store. Across six data sets it beat leading multi-round and graph-based systems by up to eleven percent on average. Retrieval costs fell by up to 45 percent at the same time. The authors pitch triplets as the sweet spot between raw text chunks and heavy graphs.
Catalog record
| CREATOR | Shengbo Gong, Xianfeng Tang, Carl Yang, Wei jin |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-02435