mlsanityData to 14 Oct 2026

Context Selection for Hypothesis and Statistical Evidence Extraction from Full-Text Scientific Articles

arXiv 2603.21193 ↗First posted 22 Mar 2026cs.CL cs.AI cs.DLcited by 0

Not on the map. Why?

Abstract (arXiv)

Extracting hypotheses and their supporting statistical evidence from full-text scientific articles is central to the synthesis of empirical findings, but remains difficult due to document length and the distribution of scientific arguments across sections of the paper. The work studies a sequential full-text extraction setting, where the statement of a primary finding in an article's abstract is linked to (i) a corresponding hypothesis statement in the paper body and (ii) the statistical evidence that supports or refutes that hypothesis. This formulation induces a challenging within-document retrieval setting in which many candidate paragraphs are topically related to the finding but differ in rhetorical role, creating hard negatives for retrieval and extraction. Using a two-stage retrieve-and-extract framework, we conduct a controlled study of retrieval design choices, varying context quantity, context quality (standard Retrieval Augmented Generation, reranking, and a fine-tuned retriever paired with reranking), as well as an oracle paragraph setting to separate retrieval failures from extraction limits across four Large Language Model extractors. We find that targeted context selection consistently improves hypothesis extraction relative to full-text prompting, with gains concentrated in configurations that optimize retrieval quality and context cleanliness. In contrast, statistical evidence extraction remains substantially harder. Even with oracle paragraphs, performance remains moderate, indicating persistent extractor limitations in handling hybrid numeric-textual statements rather than retrieval failures alone.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 27 of its 49 references that are in the map. First posted on arXiv 22 Mar 2026.

Builds on

27 of 49 references.

References · 27

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  3. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  4. 2020Longformer: The Long-Document Transformercited by 3,413
  5. 2019SciBERT: A Pretrained Language Model for Scientific Textcited by 1,414
  6. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,375
  7. 2023Large Language Models Can Be Easily Distracted by Irrelevant Contextcited by 841
  8. 2019ERASER: A Benchmark to Evaluate Rationalized NLP Modelscited by 530
  9. 2023Making Retrieval-Augmented Language Models Robust to Irrelevant Contextcited by 369
  10. 2018Multi-Task Identification of Entities, Relations, and Coreference for Scientific Knowledge Graph Constructioncited by 322
  11. 2025PaperBench: Evaluating AI's Ability to Replicate AI Researchcited by 321
  12. 2023Effective Long-Context Scaling of Foundation Modelscited by 303
15 more
  1. 2024The Power of Noise: Redefining Retrieval for RAG Systemscited by 281
  2. 2019Structural Scaffolds for Citation Intent Classification in Scientific Publicationscited by 130
  3. 2017PubMed 200k RCT: a Dataset for Sequential Sentence Classification in Medical Abstractscited by 124
  4. 2024Long-Context LLMs Meet RAG: Overcoming Challenges for Long Inputs in RAGcited by 123
  5. 2020SciREX: A Challenge Dataset for Document-Level Information Extractioncited by 113
  6. 2020Growth rates of modern science: A latent piecewise growth curve approach to model publication numbers from established and new literature databasescited by 75
  7. 2020Scientific Claim Verification with VERT5ERINIcited by 47
  8. 2025The Distracting Effect: Understanding Irrelevant Passages in RAGcited by 33
  9. 2025CURIE: Evaluating LLMs On Multitask Scientific Long Context Understanding and Reasoningcited by 28
  10. 2019Claim Extraction in Biomedical Publications using Deep Discourse Model and Transfer Learningcited by 12
  11. 2021Extracting Fine-Grained Knowledge Graphs of Scientific Claims: Dataset and Transformer-Based Resultscited by 10
  12. 2025SciRerankBench: Benchmarking Rerankers Towards Scientific Retrieval-Augmented Generated LLMscited by 5
  13. 2025AcademicEval: Live Long-Context LLM Benchmarkcited by 4
  14. 2017Extracting Core Claims from Scientific Articlescited by 3
  15. 2024How Does Knowledge Selection Help Retrieval Augmented Generation?cited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024A Review on Scientific Knowledge Extraction using Large Language Models in Biomedical Sciences
  2. 2026SIN-Bench: Tracing Native Evidence Chains in Long-Context Multimodal Scientific Interleaved Literature
  3. 2025SciRAG: Adaptive, Citation-Aware, and Outline-Guided Retrieval and Synthesis for Scientific Literature
  4. 2026Diagnosing Evidence Utilization in Long-Context and Retrieval-Augmented Language Models under Matched Evidence Conditions
  5. 2025Biomedical Hypothesis Explainability with Graph-Based Context Retrieval
  6. 2025Query-driven Document-level Scientific Evidence Extraction from Biomedical Studies
  7. 2025SCALAR: Scientific Citation-based Live Assessment of Long-context Academic Reasoning
  8. 2024L-CiteEval: Do Long-Context Models Truly Leverage Context for Responding?