Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Researchers proposed RAG-CT, a defense that detects malicious queries in Retrieval-Augmented Generation (RAG) systems by analyzing their entropy and margin distributions with a score-based detection method, according to …
Researchers proposed a semantics-consistent tabular diffusion framework, described in arXiv paper 2609.16069v1, that uses large language models to extract intra-column semantics and inter-column symbolic rules from metad…
A new arXiv paper (2609.16155v1) presents a framework that fine-tunes pre-trained Large Language Models and applies Retrieval Augmented Generation (RAG) to generate synthetic time series data for manufacturing processes,…
A new arXiv paper (2609.16161v1) presents an end-to-end integer-only quantization approach and a sparse dictionary-based KV cache compression strategy that enables LLM inference on Flash compute-in-memory devices, reduci…
A position paper posted to arXiv (2609.16189v1) argues that no language-model-enabled wargame should inform planning, doctrine, policy, or crisis response without an auditable safety case, and that open-ended wargames sh…
A new arXiv review (2609.16213v1) finds that AI uplift for biological threats currently exists primarily in digital tasks, with frontier systems exceeding expert baselines on in-silico and screening-evasion benchmarks, w…
A narrative review published on arXiv identifies eight recurring ethical and privacy issues in LLM-enabled geospatial AI (GeoAI), including passive location inference from mobility traces, spatially structured bias ampli…
A new arXiv paper (2609.16245v1) introduces Metacognitive Steering, an inference-time controller that reads the cognitive regime of Kimi 2.6, a trillion-parameter mixture-of-experts model, and dynamically composes layer-…
Researchers proposed CLEAR, an agentic framework for cross-source evidence adjudication in large language models in medicine, detailed in arXiv paper 2609.16301v1. CLEAR independently generates candidate answers from thr…
Researchers introduced Blindspot, a benchmark for trajectory-level safety calibration of long-horizon tool-using LLM agents, containing 22 attack families and 35 scenarios across seven domains that yield more than 2,500 …
A new arXiv paper (2609.16338v1) introduces BITCOS, a distribution-adaptive storage layout for ternary LLMs that costs 2 − z bits per weight given a zero density z, after measuring 29 ternary models and finding zeros acc…
A new arXiv paper (2609.16454v1) finds that supervised fine-tuning (SFT) is not inherently biased toward mode collapse or its opposite, over-dispersion, in large language models. The authors derive a bias-variance decomp…
Researchers introduced MechReason, a benchmark of 12,000 question-answer pairs and 21,000 visual materials drawn from real mechanical engineering papers, to test multi-image, multi-hop reasoning in multimodal large langu…
Researchers introduced SceneBench, a benchmark of 966 photorealistic 3D scenes reconstructed with Gaussian Splatting and annotated with over 183K hierarchical nodes, built through a human-in-the-loop pipeline using rough…
Researchers proposed ReImaGin, a method that uses image generation models as a flexible visual reasoning mechanism for multimodal large language models, according to a new arXiv paper (2609.16409v1). Across six diverse v…
A new arXiv paper (2609.16592v1) introduces an end-to-end framework for generating context-specific large language model benchmark datasets by combining expert input with synthetic data generation. The framework uses a s…
Researchers introduced VideoMM, an adaptive macro-micro inference framework for video multimodal large language models that decouples region selection from detailed reasoning, according to an arXiv paper (arXiv:2609.1672…
A new arXiv paper (2609.16589v1) reports that large language models possess an intrinsic value system, based on projecting responses from 106 LLMs at 150,000 queries per model and 95,000 human survey profiles into a shar…
A new arXiv paper (2609.16564v1) proposes a pre-generation triage layer for retrieval-augmented generation (RAG) pipelines that treats a source's material relationship to the query as query dependent, routing canonical q…
A research paper posted to arXiv introduces CoAdapt, an adaptive collaborative perception framework for Industrial IoT robotic swarms in which a Large Language Model acts as a runtime fusion controller that decides which…