Retrieval-augmented generation

Most-superseded baselines

Ranked by how many distinct papers critique or beat each method. These are the standard baselines newer work routinely measures against.

1. 1

   RAG· RAG

   RAG+: Enhancing Retrieval-Augmented Generation with Application-Aware Reasoning

   69 papers critique it · 79 beat it on benchmarks

2. 2

   GraphRAG· RAG

   A Survey of Graph Retrieval-Augmented Generation for Customized Large Language Models

   38 papers critique it · 31 beat it on benchmarks

3. 3

   Self-RAG· Self-RAG

   Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection

   26 papers critique it · 40 beat it on benchmarks

4. 4

   LightRAG· RAG

   15 papers critique it · 24 beat it on benchmarks

5. 5

   RAPTOR· RAG

   RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval

   7 papers critique it · 21 beat it on benchmarks

6. 6

   FLARE· Self-RAG

   Active Retrieval Augmented Generation

   11 papers critique it · 14 beat it on benchmarks

7. 7

   PoisonedRAG· PoisonedRAG

   PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models

   12 papers critique it · 13 beat it on benchmarks

8. 8

   HippoRAG 2· RAG

   From RAG to Memory: Non-Parametric Continual Learning for Large Language Models

   1 papers critique it · 23 beat it on benchmarks

9. 9

   IRCoT· RAG

   4 papers critique it · 19 beat it on benchmarks

10. 10

    BM25· BM25

    6 papers critique it · 16 beat it on benchmarks

11. 11

    CRAG· Self-RAG

    Corrective Retrieval Augmented Generation

    11 papers critique it · 11 beat it on benchmarks

12. 12

    ReAct· RAG

    ReAct: Synergizing Reasoning and Acting in Language Models

    6 papers critique it · 15 beat it on benchmarks