arXiv:2609.24106v1 Announce Type: new Abstract: Questions scraped from the web are used across academia and industry as a proxy for what people want to know. Across QA training data, retrieval benchm...
The paper introduces a method to detect which web scrapers feed data to large language models (LLMs) by deploying dynamic websites that issue unique canary tokens to each scraper. By querying LLMs for information about these sites, the authors can identify when an LLM consistently outputs the unique tokens, indicating exposure to a specific scraper. Experiments on 22 production LLM systems show the technique reliably uncovers both known and undisclosed scrapers, offering a tool for third parties to monitor and control unwanted web scraping.
The study investigates how often web agents incorrectly treat an HTTP 200 response as a successful fetch of useful content. By sampling 10,307 URLs—including random paths and expert‑verified claim citations—the authors found that roughly 20–25% of 200 responses lack the page’s main content, leading to inflated estimates of claim verifiability and citation drift. They propose simple heuristics and classifiers to detect unusable responses, achieving high recall with low false‑alarm rates, and recommend that agents report fetch outcomes explicitly.
The study measures Google AI Overviews (AIOs), noting they activate for 13.7% of queries overall and 64.7% of question-form queries, with lower rates for politically sensitive topics. AIOs cite more credible domains than typical search results, yet 30% of cited pages are not shown in the first-page results, indicating a distinct source selection process. Among 98,020 atomic claims, 11.0% lack support from cited pages, mainly due to omissions, and over half of cited pages contain display ads, reducing publisher revenue while Google’s ads remain.
DeGenTWeb is a system that identifies websites whose informational prose is largely machine‑generated text (MGT) with minimal human input. By adapting MGT detectors for web pages and aggregating results across sites, the study reports that 20.9% of Common Crawl‑archived sites (Jan–Jul 2026) and 13.6% of Bing‑linked sites (2025) are MGT‑dominant, with a 0.12% mislabeling rate for human‑authored sites. The prevalence of MGT‑dominant sites is increasing over time and many share the same owner.
The paper introduces concept provenance, a framework that categorizes query concepts into backstory-supported, human-central, human-tail, and candidate answer-side zones to detect knowledge intrusion in LLM-generated search queries. Applying this framework to 77,004 queries across 100 topics, 8 LLMs, and 5 prompt conditions, the authors find that 7.40% of non-generic concepts are candidate answer-side and appear in 97 of 100 topics, with human validation revealing 68.2% relaxed precision and equal contributions from knowledge and deployment intrusion. The study shows that while prompt conditions do not eliminate intrusion, post-generation concept-provenance selection can achieve 99% elimination, positioning concept provenance as a diagnostic tool rather than an evaluation predictor.
arXiv:2608. 16824v1 Announce Type: new Abstract: Generative Engine Optimization (GEO) modifies web content to increase its likelihood of being selected and cited by generative search engines.
The paper evaluates how modern large language models use internal web search to answer factual questions. Using 783 static queries and 288 dynamic queries, the authors find that enabling retrieval improves accuracy on static questions but hurts confidence calibration. On dynamic queries, models often retrieve but still achieve less than 70% accuracy, mainly due to poor query formulation and source selection, indicating that internal web search works better as a quick verification tool than a full information‑retrieval system.
The paper introduces ElephantBench, a closed‑book knowledge probe with 1,094 multi‑account factual questions generated via an auditable graph‑based pipeline that pulls documents from a low‑exposure web corpus and identifies naturally occurring disagreements. Across 32 large language models, even the best model only recovers both divergent accounts on 52.4% of questions, and most models recall one account while omitting the other, indicating persistent epistemic myopia. The study shows that scaling model size and inference‑time reasoning improves recall but does not eliminate incompleteness, and that exposure imbalance in the corpus biases models toward the dominant account.
arXiv:2606. 08649v1 Announce Type: cross Abstract: Forensic analysis of web server logs demands both accurate detection and human-readable explanations that can satisfy legal requirements.
The study audits large language model (LLM) outputs by measuring how well repeated queries recover a collected set of responses versus the full set of possible outputs. Using sample-based rarefaction on 4,500 responses from 50 buying questions across six configurations, the authors find historical-dictionary median recovery rates between 92.6% and 95.2%, which drop to 89.5%–94.7% after re‑adjudicating all candidate strings. Additional analyses with Gemini 3.1 Pro annotations and matched roster data confirm that recovery percentages vary with extraction methods, question selection, and the finite reference collection, underscoring the need for explicit measurement definitions and sensitivity analyses in LLM audits.
arXiv:2608.21829v1 Announce Type: cross Abstract: Retrieval-augmented generation treats the document store as a frozen input, and the systems that instead let an agent curate one never measure what c...