AtomCite is an agentic framework that verifies and corrects page‑level citations in multi‑page documents by parsing answers into claims, checking each claim against the cited page image, and applying a deterministic repair policy. The authors introduce DocCite, the first benchmark for this task, built on MP‑DocVQA and DUDE, containing 928 injected instances and 1,909 verified natural errors. Across Gemini, Claude, and GPT models, AtomCite achieves about 93% verification accuracy and improves citation precision from 34% to 87‑90%, while also enhancing hallucination detection in open‑source models.
By Chen Qian, Yimeng Wang, Yu Chen, Lingfei Wu, Andreas Stathopoulos
arXiv:2603. 26791v3 Announce Type: replace-cross Abstract: Assessing a cited paper's impact is typically done by analyzing its citation context in isolation within the citing paper.
By Hannah Collison, Benjamin Van Durme, Daniel Khashabi
Large language models (LLMs) are increasingly used for scientific hypothesis generation. However, evaluating generated hypotheses remains a challenge for trustworthy AI-enabled scientific workflows.
arXiv:2606.22342v2 Announce Type: replace
Abstract: How does research evolve, and can we trace it at the level of individual claims? Scientific progress is not simply a uniform accumulation of facts....
By Abdul Muntakim, Md Abdullah Al Hafiz Khan, Sadid Hasan, Yong Pei
R2VC is a modular fact‑checking system that separates retrieval, reasoning, verification, and confidence calibration. It uses hybrid sparse‑plus‑dense Wikipedia retrieval, a fine‑tuned generator for structured verdicts, an NLI cross‑encoder for selecting evidence‑based candidates, and a lightweight calibrator for confidence and abstention. On the FEVER benchmark, R2VC improves accuracy by 13.74% over a baseline and shows that verifier‑based candidate selection and calibration are key contributors to performance.
By Dhruv Dixit, Paritosh Pandey
The paper investigates whether large language models (LLMs) can reliably assess scientific hypotheses by using a logit-based energy scoring method that leverages the model’s intrinsic confidence. Across 1,323 papers in 12 disciplines, this intrinsic scoring achieved a 33.0% Hit@1 rate, outperforming a prompted listwise ranking approach that scored 16.6%. The best result, a 1‑billion‑parameter model with logit-based energy scoring, reached 53.1% Hit@1, suggesting that confidence‑based evaluation could improve trustworthy AI‑enabled scientific discovery.
By Swati Rajwal, Sanjay Das, Tirthankar Ghosal
SHELF is a Python system that creates synthetic, controlled benchmark data for evaluating large language models on bibliographic tasks such as classification, clustering, retrieval, pair classification, and instruction retrieval. It generates 62,899 model-written documents based on Library of Congress vocabularies and compares methods like TF, TF‑IDF, BM25, popular encoders, and zero‑shot decoders, reporting performance metrics such as 0.8887 for subject classification and 0.2605 for genre‑form classification. The tool also allows independent variation of bibliographic facets and can produce unseen documents beyond a model’s training cutoff, with results indicating that model rankings transfer more reliably than absolute scores when compared to other benchmarks.
By Michael J. Bommarito II
arXiv:2606. 26449v1 Announce Type: cross Abstract: Retrieval-augmented systems routinely present citations alongside generated answers, yet a citation does not confirm that the corresponding source meaningfully shaped the output.
By Mohammad Faizan, Dalal Alharthi
The paper introduces REASONS, a benchmark comprising 12,723 sentence-level citation instances across 12 arXiv subject categories, to evaluate scientific citation attribution by large language models. It proposes a dual-metric framework—Abstention Rate (AR) and Hallucination Rate (HR)—to assess the trade-off between reliability and responsiveness. Experiments on proprietary and open-source LLMs under various prompting and retrieval settings show that advanced Retrieval-Augmented Generation (RAG) reduces hallucinations but may increase abstention, while retrieval-augmented variants often maintain near-zero abstention. Human evaluation reveals a high ratio of factual hallucinations to acceptable paraphrases, underscoring the need for systems that can appropriately abstain under uncertainty.
By Deepa Tilwani, Yash Saxena, Seyedali Mohammadi, Ankur Padia, Edward Raff, Amit Sheth, Srinivasan Parthasarathy, Manas Gaur
arXiv:2602. 20459v2 Announce Type: replace Abstract: Can AI systems trained on the existing scientific record forecast the advances that will follow?
By Anirudh Ajith, Amanpreet Singh, Jay DeYoung, Nadav Kunievsky, Austin C. Kozlowski, Oyvind Tafjord, James Evans, Daniel S. Weld, Tom Hope, Doug Downey
SHELF is a Python system that creates controlled benchmark data and evaluation tasks for libraries and archives, using labelled taxonomies, writing specifications, and a generation budget. It generates 62,899 model-written documents based on Library of Congress vocabularies and supports tasks such as classification, clustering, retrieval, pair classification, and instruction retrieval. The release compares various methods—including TF, TF-IDF, BM25, popular encoders, and zero-shot decoders—showing that sparse methods remain competitive on classification and that SHELF can vary bibliographic facets independently while generating new, verifiably unseen documents.
The paper "Science or Slop?: Benchmarking and Mitigating Scientific Slop in AI-Generated Papers" introduces SciSlopBench, a dataset of 390 AI‑generated papers paired with human‑written counterparts, and defines six measures across Structure, Argument, and Artifacts to detect scientific slop. The authors show that these measures can identify AI papers with 85.9% accuracy and that higher slop correlates with lower ICLR ratings and distinguishes rejected from accepted papers. They also propose SciSlopHarness, a framework that guides a fixed LLM to revise only evidence‑supported sections, reducing the AI‑human gap by 63% without human reference targets.
By Yerim Oh, Young-Jun Lee, Jaewoo Ahn, Gunhee Kim, Dongyeop Kang