Research map / Large language models / Retrieval and knowledge
Long-context and personalized retrieval
Combine retrieval with long-context handling, calibration, and personalization benchmarks.
61 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2026) · an approach in Retrieval and knowledge, keyword cluster “retrieval augmented · augmented generation · generation rag”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
20 papers in 2023–24 and 40 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 8.8% to 6.1%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2020 · Interactive Rare-Category-of-Interest Mining from Large Datasets (AAAI 2020 · 1 citation)
- 2023 · Augmenting Language Models with Long-Term Memory (NeurIPS 2023 · 33 citations)
- 2024 · HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models (NeurIPS 2024 · 105 citations)
- 2025 · RATT: A Thought Structure for Coherent and Correct LLM Reasoning (AAAI 2025 · 22 citations)
- 2026 · DeepRAG: Thinking to Retrieve Step by Step for Large Language Models (ICLR 2026 · 3 citations)
Earliest papers here
- Interactive Rare-Category-of-Interest Mining from Large Datasets (AAAI 2020 · 1 citation)
- Augmenting Language Models with Long-Term Memory (NeurIPS 2023 · 33 citations)
- Why do Nearest Neighbor Language Models Work? (ICML 2023 · 6 citations)
What has been tried
35 papers here report 393 measured changes (a component added, removed or swapped, compared with a baseline): helped 76%, no clear effect 6%, hurt 16%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add node specificity | MuSiQue · R@5 | 50.2 -> 51.9 | Helped | HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models (NeurIPS 2024 · 105 citations) |
| Add QLoRA fine-tuning | LLeQA test · METEOR | 11.6 -> 19.7 | Helped | Interpretable Long-Form Legal Question Answering with Retrieval-Augmented Large Language Models (AAAI 2024 · 81 citations) |
| Combine the proposed scores using Logistic Regression | FAVA-Annotation · detection performance | No clear effect | LLM-Check: Investigating Detection of Hallucinations in Large Language Models (NeurIPS 2024 · 41 citations) | |
| increase RoPE base from 1.0×10^6 to 1.0×10^7 during In2 training | VaL Probing All (20% training data) · Avg | 80.3 -> 84.9 | Helped | Make Your LLM Fully Utilize the Context (NeurIPS 2024 · 40 citations) |
| Add activity patterns | Normal trajectory, normal data · DARD | 0.223 -> 0.125 | Helped | Large Language Models as Urban Residents: An LLM Agent Framework for Personal Mobility Generation (NeurIPS 2024 · 36 citations) |
| remove the manager and have the last worker generate the final response | MuSiQue · F1 | 37.09 -> 26.79 | Hurt | Chain of Agents: Large Language Models Collaborating on Long-Context Tasks (NeurIPS 2024 · 27 citations) |
| Increase generation steps from 3 to 5 | Game of 24 · success rate | 25.9 -> 31.8 | Helped | RATT: A Thought Structure for Coherent and Correct LLM Reasoning (AAAI 2025 · 22 citations) |
| Replace BM25 with embedding-based retrieval | ArcadeQA · accuracy | 37.7 -> 49.2 | Helped | TableRAG: Million-Token Table Understanding with Language Models (NeurIPS 2024 · 17 citations) |
| Remove retrieval | Ablation study · Citation recall | 83.48 (± 5.05) -> 60.11 (± 6.42) | Hurt | AutoSurvey: Large Language Models Can Automatically Write Surveys (NeurIPS 2024 · 16 citations) |
| Use Dragon instead of no retrieval | Seven long-context tasks · average score | 31.61 -> 35.73 | Helped | Retrieval meets Long Context Large Language Models (ICLR 2024 · 15 citations) |
| Add RAG | SWC-101 · F1 (%) | 79.71 -> 95.45 | Helped | SCALM: Detecting Bad Practices in Smart Contracts Through LLMs (AAAI 2025 · 11 citations) |
| Use sentence-based retrieval instead of 512-token chunk retrieval | BABILong QA1 · accuracy | Helped | BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack (NeurIPS 2024 · 8 citations) |
What reviewers will expect
Datasets and baselines named in the 21 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- NaturalQuestions · 3 of 21 papers
- HotpotQA · 3 of 21 papers
- HumanEval · 2 of 21 papers
- TriviaQA · 2 of 21 papers
Baselines
- GPT-4 · 2 of 21 papers
- ChatGPT · 2 of 21 papers
- Self-Rag · 2 of 21 papers
- Claude-3.5-Sonnet · 2 of 21 papers
- In-Context RAG · 2 of 21 papers
- GPT-4o · 2 of 21 papers
- RAPTOR · 2 of 21 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 60 papers; from OpenAlex and the papers' affiliations).
- Microsoft · 4 papers
- Renmin University of China · 3 papers
- Carnegie Mellon University · 3 papers
- Korea Advanced Institute of Science and Technology · 3 papers
- Google · 3 papers
- Alibaba Group · 3 papers
- University of Michigan - Ann Arbor · 2 papers
- Stevens Institute of Technology · 2 papers
- Harvard University · 2 papers
- Pennsylvania State University · 2 papers
Use it
Download BibTeX of 22 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Retrieval-augmented reasoning (130 papers)
- RAG robustness and evaluation (164 papers)
Elsewhere in the topic
- Knowledge graph reasoning with LLMs (174 papers)
- Knowledge editing in LLMs (170 papers)
- Explanatory and multiple-choice QA benchmarks (152 papers)
- Knowledge-augmented question answering (134 papers)
- Domain-specific QA datasets (79 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).