Method Drift›Retrieval-augmented generation
DRAGIN
DRAGIN: Dynamic Retrieval Augmented Generation based on the Information Needs of Large Language ModelsRetrieval-augmented generation · first seen Mar 15, 2024
superseded — cited as a baseline and beaten by newer methods
5 papers critique it · 8 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites DRAGIN as a baseline.
“Adaptive methods such as FLARE~jiang2023active, Self-RAG~asai2024selfrag, and DRAGIN~su2024dragin dynamically trigger retrieval based on uncertainty signals, but do so reactively, first detecting uncertainty and then blocking generation to perform retrieval.”
— Predictive Prefetching for Retrieval-Augmented Generation“These signals can offer insights into the model's reasoning process, but often require access to model weights and, consequently, cannot be used with closed-source LLMs. Additionally, the lack of interpretability in latent representations undermines trustworthiness, making these signals less suitable for applications such as healthcare, where trustworthiness plays a critical role.”
— Knowing You Don't Know: Learning When to Continue Search in Multi-round RAG through Self-Practicing“This demonstrates that relying on a single token's confidence as the signal can result in ineffective intervention, allowing multiple low-confidence tokens to be generated before retrieve.”
— Modeling Uncertainty Trends for Timely Retrieval in Dynamic RAG“However, current dynamic RAG methods fail to predict whether the LLM has the capability to answer a question prior to generation, thereby triggering retrieval in advance. Moreover, most methods often rely on static rules, leading to ineffective timing for retrieval triggers during the generation process.”
— DioR: Adaptive Cognitive Detection and Contextual Retrieval Optimization for Dynamic Retrieval-Augmented Generation“However, the former approach has limitations as LLMs tend to be overconfident, generating high-confidence probability distributions even when lacking relevant knowledge.”
— ICA-RAG: Information Completeness Guided Adaptive Retrieval-Augmented Generation for Disease Diagnosis
Beaten on benchmarks
Head-to-head results where a newer method reports beating DRAGIN. Values are copied from the source paper's tables — verify against the cited paper.
- Predictive Prefetching for Retrieval-Augmented Generation
predictive prefetching framework beats DRAGIN · F1 [HotpotQA]
75.1 vs 73.2
- Predictive Prefetching for Retrieval-Augmented Generation
predictive prefetching framework beats DRAGIN · E2E [HotpotQA]
5.2 vs 6.4
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Gemma-2b in-domain HotpotQA]
39.4 vs 22.6
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Gemma-2b in-domain NQ]
35.0 vs 22.2
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Gemma-2b in-domain TriviaQA]
52.2 vs 47.0
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Gemma-2b out-of-domain MuSiQue]
8.8 vs 4.8
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Gemma-2b out-of-domain 2Wiki]
43.6 vs 28.8
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Mistral-7b in-domain HotpotQA]
38.6 vs 28.0
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Mistral-7b in-domain NQ]
39.4 vs 37.2
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Mistral-7b in-domain TriviaQA]
52.2 vs 42.2
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Mistral-7b out-of-domain MuSiQue]
9.8 vs 7.2
- Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval
Probing-RAG beats DRAGIN · ACC [Mistral-7b out-of-domain 2Wiki]
33.4 vs 25.8
Newer alternatives
Recent methods in the same sub-problem, not yet superseded in the knowledge base.
- May 26, 2026
- May 18, 2026
- ConflictRAGConflictRAG: Detecting and Resolving Knowledge Conflicts in Retrieval Augmented GenerationMay 17, 2026
- SEMA-RAGSEMA-RAG: A Self-Evolving Multi-Agent Retrieval-Augmented Generation Framework for Medical ReasoningMay 16, 2026
- PyRAGRetrieval is Cheap, Show Me the Code: Executable Multi-Hop Reasoning for Retrieval-Augmented GenerationMay 13, 2026
- CoRM-RAGBeyond Semantic Relevance: Counterfactual Risk Minimization for Robust Retrieval-Augmented GenerationMay 2, 2026
- STEMSTEM: Structure-Tracing Evidence Mining for Knowledge Graphs-Driven Retrieval-Augmented GenerationApr 24, 2026
- Apr 22, 2026
- Self-Correcting RAGSelf-Correcting RAG: Enhancing Faithfulness via MMKP Context Selection and NLI-Guided MCTSApr 12, 2026
- Mar 7, 2026
- Cooperative Retrieval-Augmented Generation (CoRAG)Rethinking Retrieval-Augmented Generation as a Cooperative Decision-Making ProblemFeb 21, 2026
- Jan 29, 2026