publication and attribution

Traditional academic journals conflate dissemination, quality certification, attention allocation, and professional credit assignment, yet suffer from inefficiency, slow peer review, bias toward established authorities, and high costs. This work proposes decoupling dissemination from certification: authors deposit manuscripts in open repositories and undergo continuous post-publication review through signed, reply-enabled public commentary. By integrating version control and a forkable document architecture, the system automatically tracks the evolution of ideas and attributes contributions accurately. This approach constitutes the first comprehensive publishing paradigm that systematically combines open archiving, identity-transparent review, and forking mechanisms as a viable alternative to conventional journals. It ensures transparency, traceability, and fairness while effectively addressing key challenges such as low participation, chilling effects, signal dilution, and inequitable attention distribution.

This study addresses the ambiguity and lack of quantifiable assessment in author contribution allocation within scientific collaboration. Leveraging author contribution statements from over 400,000 papers, we construct the first large-scale computational framework for mapping free-text contributions to the 14 standardized CRediT roles. Our analysis reveals a significant gradient association between author position and task type: early-positioned authors predominantly perform experimental and analytical tasks, whereas last-positioned authors concentrate on leadership and management responsibilities. Within small teams, individual task loads vary by over threefold, with disparities scaling linearly with team size. We standardize 5.6 million author–task assignments across 1.58 million author mentions. This constitutes the first empirical demonstration that contemporary scientific collaboration exhibits a “position-driven” center–periphery division of labor and a hierarchical role stratification.

Individual contributions in scientific teams remain difficult to identify, as conventional author ordering and self-reported statements suffer from bias and incomplete coverage. Method: Leveraging LaTeX source files from 1.6 million papers (1991–2023), we introduce a large-scale, code-level analysis of implicit labor division in scientific writing—distinguishing conceptual (e.g., Introduction, Discussion) from technical (e.g., Methods, Experiments) authorship. We innovatively treat author-defined macros as objective, fine-grained writing traces to quantify chapter-level contributions, bypassing reliance on positional heuristics or subjective declarations. Contribution/Results: Integrating author-chapter mapping, multi-source validation (CRediT statements, Overleaf logs, disciplinary norms), and statistical testing (Spearman’s ρ = 0.6, *p* < 0.05), our method achieves 0.87 accuracy. We release the first global, million-scale dataset of scientific writing contributions—revealing robust cross-disciplinary division-of-labor patterns and providing a scalable, empirical foundation for equitable author credit allocation and research evaluation reform.

Scientific software development remains systematically under-recognized in academic credit allocation, creating a misalignment between code contributions and traditional authorship. Method: Leveraging ~140,000 paper–code repository pairings, we constructed a cross-platform author–developer linkage prediction model and conducted multivariate statistical analyses controlling for confounding factors. Contribution/Results: We find that nearly 30% of papers have code contributors who are not listed as authors; these contributors are rarely acknowledged, and their contributions yield no significant increase in paper citations. Crucially, researchers with high coding activity exhibit significantly lower h-indices. These findings demonstrate that current scholarly evaluation systems substantially undervalue software contributions, resulting in a misalignment between incentive structures and research practice. This study provides the first large-scale empirical evidence of a “recognition deficit” for software labor, offering critical support for reforming research assessment frameworks.

This study addresses the risk of information loss and disruption to the scholarly record posed by the absence of long-term preservation mechanisms for academic blogs, an emerging form of scholarly communication. Employing a convergent mixed-methods approach, the research integrates quantitative analysis of 866 German-language academic blogs, in-depth interviews with 13 bloggers, and open community-based participatory review to systematically synthesize multi-source evidence for the first time. The work proposes a comprehensive set of digital preservation requirements specifically tailored to academic blogs and develops an actionable implementation guide for library practitioners. This framework offers an innovative solution to effectively integrate academic blogs into scholarly information infrastructures, ensuring their long-term accessibility, reusability, and citability.

This work addresses persistent issues in academic manuscripts—such as erroneous citation identifiers, missing metadata, misattributed authorship, and confusion between preprints and published versions—exacerbated by the propensity of large language models to generate citation hallucinations. To mitigate these challenges, we propose a TypeScript-based Model Context Protocol (MCP) server that integrates automated citation validation into intelligent scholarly editing workflows for the first time. Our system employs a manifestation-aware matching mechanism and policy-gated rewriting strategies, harmonizing data from multiple sources including PubMed, Crossref, arXiv, and Semantic Scholar. It supports structured parsing of diverse file formats and multi-round retrieval to generate precise correction suggestions. The prototype has been rigorously evaluated across 47 test cases covering repair actions, exception handling, and protocol compliance, demonstrating robust defense against both conventional citation errors and LLM-induced hallucinations.

This work addresses the critical challenge of hallucinated citations in scientific texts generated by large language models, which pose a serious threat to academic integrity and are difficult to detect manually. To this end, the authors introduce the first benchmark and verification framework specifically designed for detecting fabricated references in scientific writing. The framework features a novel unified metric for evaluating citation faithfulness and evidence alignment, supported by a large-scale, cross-domain dataset rigorously validated by human annotators. Methodologically, it proposes an interpretable and scalable multi-agent verification pipeline that integrates claim extraction, evidence retrieval, passage matching, and reasoning calibration to enable end-to-end auditing of citation authenticity. Experimental results demonstrate that the proposed approach significantly outperforms existing methods in both accuracy and interpretability, effectively identifying citation errors produced by state-of-the-art large language models and offering a reliable tool for scientific publishing integrity.

This study addresses the widespread yet often unintentional leakage of sensitive information—such as private URLs, API keys, and Git history—in arXiv preprint source files, posing significant risks to author privacy and security. Conducting the first large-scale quantitative analysis of 2.7 million arXiv submissions, we systematically identify leakage vulnerabilities across three dimensions: redundant files, embedded metadata, and irrelevant code comments. To mitigate these risks, we introduce ALC-NG, the first comprehensive sanitization tool capable of simultaneously cleaning auxiliary files, metadata, and source-code annotations. ALC-NG integrates automated static analysis, metadata extraction, textual pattern recognition, and LaTeX compilation validation. Our experiments reveal that nearly all submissions contain hidden sensitive data, that existing tools offer limited protection, and that ALC-NG effectively removes unnecessary content while preserving successful document compilation.

This work addresses the pervasive issue of citation errors in scientific literature, which existing methods struggle to verify at scale due to reliance on abstracts or limited datasets. We propose BibAgent, an end-to-end agent framework that integrates large language models, cross-source document retrieval, and an adaptive evidence aggregation mechanism, employing tailored strategies for open-access and paywalled publications. A key innovation is the Evidence Committee mechanism, which infers the validity of citations to restricted-content papers through consensus among downstream citing works. To support systematic evaluation, we introduce MisciteBench—a large-scale, cross-disciplinary benchmark comprising 6,350 samples spanning five categories of miscitation. Experiments demonstrate that BibAgent significantly outperforms existing LLM-based baselines in both accuracy and interpretability, enabling efficient, traceable, and scalable detection of citation errors.

This study addresses the prevalence of unreliable citations in in-depth research reports generated by large language models (LLMs), which often suffer from broken links, irrelevant content, or factual inaccuracies. The authors propose the first end-to-end verifiable evaluation framework specifically designed for LLM-generated citations: it extracts inline references from Markdown reports via abstract syntax tree (AST) parsing, retrieves source content using retrieval-augmented generation (RAG), and conducts a closed-loop assessment across three dimensions—link validity, content relevance, and factual accuracy. The framework reveals a significant disconnect between superficial citation quality and factual reliability: state-of-the-art models achieve over 94% link validity and more than 80% relevance, yet their factual accuracy ranges only from 39% to 77%. Moreover, deeper research synthesis correlates with an average 42% decline in factual accuracy. The work also releases a large-scale citation verification pipeline and human-calibrated scoring criteria.