References

Full citations with links for every paper referenced in this repo, grouped by topic. Each entry includes a one-line note on why (and when) to read it.

Reading strategy: you don't need to read papers to pass most RAG interviews — the Q&A files summarize what matters. Read the Foundations papers if you're interviewing for research-adjacent roles; skim the architecture papers for whichever RAG types your target role emphasizes.

Foundations

Paper Link Why read it
Lewis et al., 2020 — Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks arXiv:2005.11401 The original RAG paper; defines retrieve-then-generate. Used in 02_interview_bank/01-naive-rag.md
Karpukhin et al., 2020 — Dense Passage Retrieval for Open-Domain Question Answering arXiv:2004.04906 DPR — the bi-encoder dense retrieval recipe everything else builds on
Reimers & Gurevych, 2019 — Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks arXiv:1908.10084 Why bi-encoders exist; the sentence-transformers library behind most embedding fine-tuning
Gao et al., 2023 — Retrieval-Augmented Generation for Large Language Models: A Survey arXiv:2312.10997 The survey that popularized the Naive → Advanced → Modular taxonomy used in this repo

Architecture papers (by interview bank section)

Paper Link Bank section
Yao et al., 2022 — ReAct: Synergizing Reasoning and Acting in Language Models arXiv:2210.03629 04-agentic-rag.md — the reason/act loop agentic RAG is built on
Jiang et al., 2023 — Active Retrieval Augmented Generation (FLARE) arXiv:2305.06983 04-agentic-rag.md — generate, detect uncertainty, re-retrieve
Edge et al., 2024 — From Local to Global: A Graph RAG Approach to Query-Focused Summarization arXiv:2404.16130 05-graph-rag.md — Microsoft GraphRAG: KG + community detection
Yan et al., 2024 — Corrective Retrieval Augmented Generation (CRAG) arXiv:2401.15884 06-corrective-rag.md — retrieval evaluator + web-search fallback
Asai et al., 2023 — Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection arXiv:2310.11511 07-self-rag.md — reflection tokens, trained retrieval decisions
Wang et al., 2024 — Speculative RAG: Enhancing Retrieval Augmented Generation through Drafting arXiv:2407.08223 08-speculative-rag.md — small drafter + large verifier
Radford et al., 2021 — Learning Transferable Visual Models From Natural Language Supervision (CLIP) arXiv:2103.00020 09-multimodal-rag.md — the joint text-image embedding space
Faysse et al., 2024 — ColPali: Efficient Document Retrieval with Vision Language Models arXiv:2407.01449 09-multimodal-rag.md — late-interaction retrieval over page images, no OCR
Liu et al., 2023 — Lost in the Middle: How Language Models Use Long Contexts arXiv:2307.03172 10-long-context-rag.md — why stuffing context fails; U-shaped attention
Jiang et al., 2023 — LLMLingua: Compressing Prompts for Accelerated Inference of LLMs arXiv:2310.05736 10-long-context-rag.md — prompt compression for token cost
Jeong et al., 2024 — Adaptive-RAG: Learning to Adapt Retrieval-Augmented LLMs through Question Complexity arXiv:2403.14403 11-adaptive-rag.md — query-complexity classifier routing
Trivedi et al., 2023 — Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions (IRCoT) arXiv:2212.10509 19-iterative-multihop-rag.md — interleave CoT with retrieval
Press et al., 2022 — Measuring and Narrowing the Compositionality Gap in Language Models (Self-Ask) arXiv:2210.03350 19-iterative-multihop-rag.md — explicit follow-up sub-questions
Shao et al., 2023 — Enhancing Retrieval-Augmented LMs with Iterative Retrieval-Generation Synergy (ITER-RETGEN) arXiv:2305.15294 19-iterative-multihop-rag.md — generated answer as next-round query
Gutiérrez et al., 2024 — HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models arXiv:2405.14831 20-hipporag.md — Personalized PageRank over an LLM-built KG
Packer et al., 2023 — MemGPT: Towards LLMs as Operating Systems arXiv:2310.08560 21-memory-conversational-rag.md — OS-style tiered memory paging
Gao et al., 2022 — Precise Zero-Shot Dense Retrieval without Relevance Labels (HyDE) arXiv:2212.10496 22-hyde-rag.md — embed an LLM-generated hypothetical answer
Jiang et al., 2023 — Active Retrieval Augmented Generation (FLARE) arXiv:2305.06983 23-flare-rag.md — confidence-gated, forward-looking active retrieval
Liang et al., 2024 — KAG: Boosting LLMs in Professional Domains via Knowledge Augmented Generation arXiv:2409.13731 24-kag.md — logical-form reasoning + KG/text mutual indexing
Li et al., 2024 — GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of LLMs arXiv:2406.14550 25-graphreader-gnn-rag.md — agentic graph-of-notes traversal
Mavromatis & Karypis, 2024 — GNN-RAG: Graph Neural Retrieval for Large Language Model Reasoning arXiv:2405.20139 25-graphreader-gnn-rag.md — GNN-retrieved reasoning subgraphs for KGQA
Guu et al., 2020 — REALM: Retrieval-Augmented Language Model Pre-Training arXiv:2002.08909 26-realm.md — end-to-end learned retriever via masked-LM pre-training
Borgeaud et al., 2021 — Improving Language Models by Retrieving from Trillions of Tokens (RETRO) arXiv:2112.04426 27-retro.md — chunked cross-attention over a trillion-token datastore
Izacard et al., 2022 — Atlas: Few-shot Learning with Retrieval Augmented Language Models arXiv:2208.03299 28-atlas.md — jointly-trained Contriever + FiD for few-shot knowledge tasks
Izacard & Grave, 2020 — Leveraging Passage Retrieval with Generative Models for Open Domain QA (Fusion-in-Decoder) arXiv:2007.01282 29-fusion-in-decoder.md — encode passages separately, fuse in the decoder
Jin et al., 2025 — Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning arXiv:2503.09516 42-search-r1-reasoning-rag.md — RL-trained interleaved reasoning + search tokens
Song et al., 2025 — R1-Searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning arXiv:2503.05592 42-search-r1-reasoning-rag.md — sibling RL-trained search policy
Chen et al., 2025 — ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning arXiv:2503.19470 42-search-r1-reasoning-rag.md — NeurIPS 2025
Jiang et al., 2025 — DeepRetrieval: Hacking Real Search Engines and Retrievers with LLMs via Reinforcement Learning arXiv:2503.00223 42-search-r1-reasoning-rag.md — COLM 2025
OpenAI, 2025 — Introducing Deep Research (blog) openai.com/index/introducing-deep-research 43-deep-research-rag.md — multi-step agentic research producing cited long-form reports
Qian et al., 2024 — MemoRAG: Boosting Long Context Processing with Global Memory-Enhanced Retrieval Augmentation arXiv:2409.05591 44-memorag.md — WWW 2025; global memory model generates query-time retrieval clues
Jiang et al., 2024 — LongRAG: Enhancing Retrieval-Augmented Generation with Long-context LLMs arXiv:2406.15319 45-longrag.md — large retrieval units + long-context reader
Li et al., 2024 — Retrieval Augmented Generation or Long-Context LLMs? A Comprehensive Study and Hybrid Approach (Self-Route) arXiv:2407.16833 45-longrag.md — Google; routes per-query between RAG and long-context
Yu et al., 2024 — VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documents arXiv:2410.10594 46-visrag.md — embeds and reads document pages as images, no OCR
Edge, Trinh & Larson, 2024 — LazyGraphRAG: Setting a New Standard for Quality and Cost (blog) microsoft.com/research/blog/lazygraphrag 47-lazygraphrag.md — defers all LLM summarization to query time
Wang et al., 2024 — Astute RAG: Overcoming Imperfect Retrieval Augmentation and Knowledge Conflicts for LLMs arXiv:2410.07176 48-astute-rag.md — Google; reconciles parametric and retrieved knowledge
Yu, Zhang & Feng, 2024 — Auto-RAG: Autonomous Retrieval-Augmented Generation for Large Language Models arXiv:2411.19443 49-auto-rag-deeprag.md — autonomous multi-turn retrieve decisions
Guan et al., 2025 — DeepRAG: Thinking to Retrieve Step by Step for Large Language Models arXiv:2502.01142 49-auto-rag-deeprag.md — retrieval as an MDP over atomic subqueries
Wang et al., 2025 — Chain-of-Retrieval Augmented Generation (CoRAG) arXiv:2501.14342 50-corag.md — Microsoft, NeurIPS 2025; trained retrieval chains, test-time compute scaling
Chan et al., 2024 — RQ-RAG: Learning to Refine Queries for Retrieval Augmented Generation arXiv:2404.00610 51-rq-rag.md — fine-tuned rewrite/decompose/disambiguate refinement operations
Lin et al., 2025 — REFRAG: Rethinking RAG based Decoding arXiv:2509.01092 52-refrag.md — Meta; compressed chunk-embeddings + RL-selective expansion, ~30x TTFT

Retrieval & indexing

Paper Link Why read it
Robertson & Zaragoza, 2009 — The Probabilistic Relevance Framework: BM25 and Beyond DOI:10.1561/1500000019 The definitive BM25 reference behind sparse/hybrid retrieval
Formal et al., 2021 — SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval arXiv:2109.10086 Learned sparse retrieval — the modern half of hybrid search. See 01_concepts/retrieval_strategies.md
Khattab & Zaharia, 2020 — ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction arXiv:2004.12832 Late interaction — the middle ground between bi- and cross-encoders
Gao et al., 2022 — Precise Zero-Shot Dense Retrieval without Relevance Labels (HyDE) arXiv:2212.10496 Hypothetical document embeddings for query rewriting. See 02-advanced-rag.md
Zheng et al., 2023 — Take a Step Back: Evoking Reasoning via Abstraction in LLMs arXiv:2310.06117 Step-back prompting for query abstraction. See 01_concepts/retrieval_strategies.md
Cormack et al., 2009 — Reciprocal Rank Fusion Outperforms Condorcet and Individual Rank Learning Methods DOI:10.1145/1571941.1572114 RRF — the standard way to merge dense + sparse result lists
Malkov & Yashunin, 2016 — Efficient and Robust Approximate Nearest Neighbor Search Using HNSW Graphs arXiv:1603.09320 The index behind most vector DBs. See 01_concepts/vector_databases.md
Sun et al., 2023 — Is ChatGPT Good at Search? Investigating Large Language Models as Re-Ranking Agents (RankGPT) arXiv:2304.09542 LLM listwise reranking. See 01_concepts/reranking.md
Kusupati et al., 2022 — Matryoshka Representation Learning arXiv:2205.13147 Truncatable embeddings — why text-embedding-3 lets you cut dimensions. See 01_concepts/embeddings.md
Wang et al., 2021 — GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval arXiv:2112.07577 Synthetic training data for embedding fine-tuning. See 01_concepts/fine_tuning.md
2025 — Training-Free Adaptive Retrieval Gating for Efficient RAG (TARG) arXiv:2511.09803 Training-free retrieve-or-not decision via token-entropy/logit-margin uncertainty on a no-context draft. See 02_interview_bank/11-adaptive-rag.md

Evaluation & benchmarks

Paper Link Why read it
Es et al., 2023 — RAGAS: Automated Evaluation of Retrieval Augmented Generation arXiv:2309.15217 Reference-free RAG evaluation — the de facto standard. See 01_concepts/evaluation_metrics.md
Liu et al., 2023 — G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment arXiv:2303.16634 LLM-as-judge with chain-of-thought rubrics
Thakur et al., 2021 — BEIR: A Heterogeneous Benchmark for Zero-shot Evaluation of IR Models arXiv:2104.08663 The retrieval generalization benchmark
Muennighoff et al., 2022 — MTEB: Massive Text Embedding Benchmark arXiv:2210.07316 How embedding models are compared; read before trusting any leaderboard number
Saad-Falcon et al., 2023 — ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems arXiv:2311.09476 Fine-tuned judges + prediction-powered inference for calibrated scores. See 08_evaluation/03-llm-judge-frameworks.md
Ru et al., 2024 — RAGChecker: A Fine-grained Framework for Diagnosing Retrieval-Augmented Generation arXiv:2408.08067 Claim-level entailment checking for retriever-vs-generator diagnosis. See 08_evaluation/03-llm-judge-frameworks.md
Liu et al., 2025 — Judge as A Judge: Improving the Evaluation of RAG through the Judge-Consistency of LLMs (ConsJudge) arXiv:2502.18817 Training judges via cross-judgment consistency + DPO, no human labels needed. See 08_evaluation/03-llm-judge-frameworks.md
Tamber et al., 2025 — Benchmarking LLM Faithfulness in RAG with Evolving Leaderboards (FaithJudge) arXiv:2505.04847 Human-guided (few-shot) LLM judge for hallucination, built on Vectara's Hallucination Leaderboard. See 08_evaluation/03-llm-judge-frameworks.md
Katsis et al., 2025 — MTRAG: A Multi-Turn Conversational Benchmark for Evaluating Retrieval-Augmented Generation Systems arXiv:2501.03468 Human-authored multi-turn RAG benchmark; single-turn benchmarks miss this failure surface. See 01_concepts/evaluation_metrics.md

Security

Paper Link Why read it
Greshake et al., 2023 — Not What You've Signed Up For: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection arXiv:2302.12173 Indirect prompt injection — the core RAG attack surface. See 01_concepts/prompt_injection_risks.md
Hines et al., 2024 — Defending Against Indirect Prompt Injection Attacks With Spotlighting arXiv:2403.14720 Marking untrusted content so the model can tell instructions from data
Zou et al., 2024 — PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models arXiv:2402.07867 USENIX Security 2025. Multi-document corpus poisoning to force a targeted wrong answer. See 01_concepts/prompt_injection_risks.md
Zhang et al., 2025 — Practical Poisoning Attacks against Retrieval-Augmented Generation (CorruptRAG) arXiv:2504.03957 Single-document poisoning — cheaper, stealthier variant of PoisonedRAG's threat model. See 01_concepts/prompt_injection_risks.md
Liang et al., 2025 — SafeRAG: Benchmarking Security in Retrieval-Augmented Generation of Large Language Model arXiv:2501.18636 ACL 2025. Benchmark spanning noise/conflict/injection/DoS attacks across 14 RAG components. See 01_concepts/prompt_injection_risks.md