Tag
Ai2 open-sources research report model AstaBrief 8B, built on Qwen3-8B, which can generate cited research reports in a single step. It reaches 90.5 citation precision on ScholarQA-CS2, a significant improvement over the Qwen3-8B baseline, with weights and training data released under Apache 2.0.
An individual used their OpenClaw AI agent to solve circle packing problems, resulting in 10 new packings that beat existing records, and the agent was cited on a math records page alongside professional researchers.
A fully offline AI reads over 4,000 pages of declassified UFO files using OCR and vector database, providing cited answers locally without cloud or API keys.
Elon Musk announces a /deep-research command for Grok Build that performs research with bounded parallel agents, cross-checks evidence, and generates cited reports.
A user reports that a local LLM hallucinates citations with high confidence when adapted for legal documents, and seeks advice on grounding, model, or pipeline ideas to mitigate this issue.
A new benchmark paper, OpenBioRQ, reveals that AI agents rarely fabricate citations but often cite papers that do not support the claim, with 15.9% of citations being mismatched in biomedical contexts.
A tweet highlighting the most-cited papers and most-starred GitHub repos related to ICML 2026, with publicly available code and workflows to map papers to Semantic Scholar citation data and GitHub repos.
Analysis of LLM usage in legal and compliance tasks reveals that models often produce confident but unverifiable citations, raising questions about reliable legal grounding for AI outputs.
A solo developer built Athena Wisdom, a free tool that answers questions about uploaded PDFs and other documents with explicit source citations, ensuring accuracy and transparency.
AI search engines extract and paraphrase content from pages, prioritizing citeable, evidence-backed material over generic summaries. Content creators should focus on writing for citation rather than just clicks.
Yohei Nakajima celebrates the first citation of ActiveGraph on arXiv, where it is referenced as a complementary runtime, and notes that a working bridge example has been included in the repo.
A team from NUS open-sourced PaperDebugger, a multi-agent system that lives inside Overleaf, providing real-time rewriting, critique, and citation assistance with an open enhancer model (XtraGPT-7B), making Overleaf a full research environment.
Rutgers researchers trace citation hallucination in LLMs to sparse field-specific neurons, showing causal intervention can suppress fake references.