AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- LessWrongresearchTasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts
📝 Tweet , 📄 Paper We introduce TasteVal, a benchmark that measures the experimental research taste of AI models on AI R&D tasks. We define research taste as the ability to pick interesting problems to solve, design exp…
- LessWrongresearchHow funny are the frontier AI models?
This is a crosspost from my substack , where I share results of various experiments in AI and beyond. TLDR: yes, models are getting funnier over time I love laughing. Well, who doesn’t? Good jokes have a certain notion o…
- OpenAI NewsresearchSharing AI progress in mathematics
OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.
- The DecoderresearchResearchers stretch LeCun's JEPA AI into a universal world model that works from physics to biology
Researchers at PhAI Labs have expanded Yann LeCun's JEPA architecture to work across seven fields, from robotics to biomedicine. The effort also produced a liver cancer treatment candidate that showed promise in lab test…
- LessWrongresearchCan you predict cross-propensity spillover?
As in you specifically. tldr: My SPAR group ran a 29 × 29 propensity SFT spillover experiment (i.e. how does SFT for propensity A (Ptrain) change propensity B (Ptarget) for all combinations of the 29 propensities we had…
- LessWrongresearchWhat's the Type of an Ontological Mismatch?
(Brief post, just barely not a quick take. This type signature gets put to work in Rocks to Rules , which locates mismatches on a ladder of binding-types.) Lately I've been thinking about ontological mismatches a lot - e…
- arXiv cs.AIresearchREACT: Physically and Chemically Consistent Reconstruction of Marine Active Tracers
arXiv:2610.03888v1 Announce Type: new Abstract: Reconstructing global sea surface pH from sparse observations is critical for monitoring ocean acidification and understanding marine carbon cycling. Traditional assimilati…
- arXiv cs.AIresearchSGAnalog: An End-to-End Circuit Benchmark from Open-Source Silicon Tapeouts
arXiv:2610.03934v1 Announce Type: new Abstract: Existing analog integrated circuit design benchmarks make two questions hard to answer: whether a model has learned transferable circuit skills rather than recalled familia…
- arXiv cs.AIresearchLatentQuant: Preserving the Policy-Facing Latent Contract under NVFP4 VAE Quantization
arXiv:2610.03959v1 Announce Type: new Abstract: Recent world action models (WAMs) reuse pretrained video VAEs whose encoder latents directly condition downstream action policies. Quantization must therefore preserve not…
- arXiv cs.AIresearchROAR: Unifying Runs across Heterogeneous AI-Driven Research Systems
arXiv:2610.03966v2 Announce Type: replace Abstract: Each run of an AI-driven research system (ADRS) is an expensive search over a vast solution space, and dependable evaluation requires many runs, making run data both co…
- arXiv cs.AIresearchBehavioral History Outperforms Descriptions of the Person for LLM Synthetic Personas
arXiv:2610.03998v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as synthetic personas representing survey respondents. Their validity as substitutes for particular respondents depen…
- arXiv cs.AIresearchExploration-Preserving Policy Optimization
arXiv:2610.04011v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards improves reasoning, while the allocation of learning signal shapes which solutions remain accessible under repeated sampling.…
- arXiv cs.AIresearchBeyond the Parameter Monolith: Reconstructive Memories, Executable Skills, and Residual Assembly for Language Models
arXiv:2610.04012v1 Announce Type: new Abstract: Language-model systems can separate contextual computation, persistent storage, and exact execution instead of updating all capabilities through one shared parameter system…
- arXiv cs.AIresearchA Quantitative Analysis of Graph Representation Strategies for Cyber Attack Detection
arXiv:2610.04019v1 Announce Type: new Abstract: Graph based cyber attack detection studies employ various graph construction and representation strategies across different cybersecurity application domains. This diversit…
- arXiv cs.AIresearchDiscrete Diffusion for Large Graph Generation via Structural Candidate Restriction
arXiv:2610.04056v1 Announce Type: new Abstract: Synthesizing realistic graphs at scale is vital when the graphs of interest are large and real-world samples are limited or access-sensitive. Diffusion-based generators hav…
- arXiv cs.AIresearchReinforcement Learning with Comparative Evidence for Social Intelligence
arXiv:2610.04072v1 Announce Type: new Abstract: Developing socially intelligent AI remains heavily dependent on human-annotated data, limiting the scale and breadth of social understanding models can acquire. Methods tha…
- arXiv cs.AIresearchThe Independence Prior of SAEs Fragments Visual Concepts
arXiv:2610.04112v1 Announce Type: new Abstract: Sparse Autoencoders (SAEs) decompose model activations into sparse combinations of interpretable dictionary atoms. Although SAEs are grounded in the Linear Representation H…
- arXiv cs.AIresearchLanguage Model Activations Inhabit Privileged Error-Correcting Basins
arXiv:2610.04183v1 Announce Type: new Abstract: Language models exhibit remarkable robustness, continuing to produce coherent text even when their activations are perturbed by interventions like linear steering. We hypot…
- arXiv cs.AIresearchALoDLM: Adaptively Looped Diffusion Language Models
arXiv:2610.04198v1 Announce Type: new Abstract: Diffusion language models (DLMs) enable fast generation by predicting multiple tokens in parallel, but their practical adoption remains limited by a persistent quality gap…
- arXiv cs.AIresearchFine-Tuning VLM for Enhancing AI's Spatial Intelligence: Understanding 3D and 2D Rotations
arXiv:2610.04206v3 Announce Type: replace Abstract: Spatial intelligence is a fundamental skill in multiple domains, such as Science, Technology, Engineering, and Mathematics (STEM), Medicine, Architecture, and Construct…
- arXiv cs.AIresearchTCMClinicalReason-Bench: Can Language Models Reason from Pathogenesis to Prescription over Real-World Clinical Cases?
arXiv:2610.04215v1 Announce Type: new Abstract: Large language models (LLMs) can generate clinical narratives that are insufficiently grounded in patient-specific evidence. In traditional Chinese medicine (TCM), errors c…
- arXiv cs.AIresearchOn the Steering Dimensionality of Refusal in Language Models
arXiv:2610.04245v1 Announce Type: new Abstract: Existing activation steering methods often assume that a high-level concept can be mediated by a single steering direction. To support this, two complementary interventions…
- arXiv cs.AIresearchSpec2Game: Can LLMs Generate Complete Playable Games from Detailed Specifications?
arXiv:2610.04253v1 Announce Type: new Abstract: Generating an executable program does not necessarily mean that it correctly implements the behavioral requirements specified in natural language. To evaluate large languag…
- arXiv cs.AIresearchDense Neuro-Symbolic Reasoning in a Unified Geometry State
arXiv:2610.04280v1 Announce Type: new Abstract: Geometry reasoning is naturally stateful: solving a problem repeatedly alternates between structural proposals and exact deductions. We formulate this process as dense neur…
- arXiv cs.AIresearchLanguage-Conditioned Token and Reasoning Efficiency in Large Language Models: A Paired Cross-Lingual Study Protocol
arXiv:2610.04295v1 Announce Type: new Abstract: Large language models incur language-dependent representation and inference costs, but existing comparisons often conflate input language, assigned observable-trace languag…
- arXiv cs.AIresearchEnvDreamer: Large-Scale Multimodal-to-Environment Generation for Embodied AI
arXiv:2610.04301v1 Announce Type: new Abstract: Large datasets and high capacity models have accelerated progress in vision and language. This work introduces a platform aimed at bringing comparable gains to embodied lea…
- arXiv cs.AIresearchMOIRA: Mass-Oriented Indexing with Ragged Attention for Long-Context Decoding
arXiv:2610.04313v1 Announce Type: new Abstract: Long-context decoding is limited by memory bandwidth, because every output token reads the KV cache of every layer. Sparse decoding reduces this cost by reading only part o…
- arXiv cs.AIresearchFrom Latent Space to Jacobian Space: Measuring, Evading, and Training Against Safety-Content Accessibility
arXiv:2610.04316v1 Announce Type: new Abstract: Output-only safety monitoring sees only the end of a model's computation, yet the model computes its answer before emitting it: what it is internally poised to say is safet…
- arXiv cs.AIresearchBidirectional Preference Synthesis: Learning Prompt-Conditioned Preferences from Boundary Failures
arXiv:2610.04328v1 Announce Type: new Abstract: Correction-based offline preference pipelines commonly treat model failures only as rejected responses under the original prompt. This supervision is incomplete for boundar…
- arXiv cs.AIresearchSuppressing Pressure, Amplifying Evidence: Self-Guided Attention Steering to Mitigate Sycophancy and Stubbornness
arXiv:2610.04329v1 Announce Type: new Abstract: Reliable language models should resist unsupported user pressure while effectively using objective contextual information. However, models may exhibit sycophancy by yieldin…
- arXiv cs.AIresearchShadowMiner v1 - An Experience Report on Implementing and Measuring a Problem-and-Hypothesis Discovery Engine
arXiv:2610.04339v1 Announce Type: new Abstract: ShadowMiner v1 is a system that automatically discovers research problems and generates hypotheses from AI papers. It is a nine-stage pipeline. It structures documents into…
- arXiv cs.AIresearchFunctionally Equivalent or Not? Graph-Grounded Differential Surrogate Execution for Code Equivalence
arXiv:2610.04371v1 Announce Type: new Abstract: Determining whether two programs are functionally equivalent is central to code modernization, patch validation, refactoring, and code-generation evaluation. Yet the usual…
- arXiv cs.AIresearchTrustMed-RL: Long-Horizon Reinforcement Learning for Evidence-Grounded Clinical Diagnosis
arXiv:2610.04387v1 Announce Type: new Abstract: Medical language models can produce correct diagnoses despite incomplete investigations and unsupported reasoning. To support long-horizon, evidence-grounded diagnosis, we…
- arXiv cs.AIresearchTimeNet: An Extensible Unified Data Infrastructure for Next-Generation Temporal Foundation Models
arXiv:2610.04407v1 Announce Type: new Abstract: Temporal Foundation Models (TFMs) aim to generalize across domains, datasets, and tasks. Yet, their development remains constrained by fragmented, task-specific data format…
- arXiv cs.AIresearchRAGrasp: Geometry-Semantic Template Retrieval and Grasp Transfer
arXiv:2610.04438v1 Announce Type: new Abstract: We present RAGrasp, a retrieval-augmented pipeline for planar parallel-jaw grasping from a compact set of locally collected, grasp-annotated RGB-D (color and depth) templat…
- arXiv cs.AIresearchCausally Fair Generation with Large Language Models
arXiv:2610.04444v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to generate, complete, and transform information in settings where their outputs can shape consequential decisions, raisi…
- arXiv cs.AIresearchReactivating Alignment: Defending LLMs from Jailbreaks via Intention-Aware Input-Output Matching
arXiv:2610.04470v1 Announce Type: new Abstract: Large language models (LLMs) remain vulnerable to jailbreak attacks that conceal harmful intent within complex adversarial prompts. Existing defenses primarily rely on inpu…
- arXiv cs.AIresearchSemantic Causal-Factor Inference from Aviation Incident Narratives Using A Variational Autoencoder with Cosine-Similarity-Based Reconstruction
arXiv:2610.04472v1 Announce Type: new Abstract: Aviation accident and incident investigations generate extensive unstructured textual information containing evidence relevant to the causes and contributing factors of saf…
- arXiv cs.AIresearchInferOpt: Constrained Multi-Objective Search for LLM Inference Configurations
arXiv:2610.04473v1 Announce Type: new Abstract: Serving an LLM means setting dozens of inference-time knobs, from per-layer KV retention to per-layer expert counts. Practice sets them with mechanism-specific heuristics t…
- arXiv cs.AIresearchVCLMU: Mechanism-Centric Virtual Cell World Modeling for Perturbation Response
arXiv:2610.04475v1 Announce Type: new Abstract: Predicting cellular responses to genetic perturbations is a central capability for virtual cells and a key step toward computational modeling of biological interventions. M…
- arXiv cs.AIresearchManifoldCache: Training-Free Diffusion Acceleration via Constraint Manifold Caching
arXiv:2610.04510v2 Announce Type: replace Abstract: Diffusion models for structured scientific generation must produce samples satisfying hard geometric constraints imposed by physics, chemistry, or biology, yet inferenc…
- arXiv cs.AIresearchAction-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models
arXiv:2610.04539v1 Announce Type: new Abstract: Latent world models learn to predict observed transitions, yet low prediction error alone does not guarantee reliable planning. Inspired by self tickling experiments in neu…
- arXiv cs.AIresearchRecursive Improvement of a Differentiable Scientific Software Ecosystem
arXiv:2610.04561v2 Announce Type: replace Abstract: Differentiable programming connects scientific computation with gradient-based inference, learning and design. Extending these capabilities across a heterogeneous softw…
- arXiv cs.AIresearchStrong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs
arXiv:2610.04580v1 Announce Type: new Abstract: Multi-modal large language models (MLLMs) achieve strong modality understanding by pairing a large language model (LLM) with an encoder for a target modality such as vision…
- arXiv cs.AIresearchBounds, Decompositions and Null Behaviour of KRATOS: A Mathematical Specification of a Recognition-Comparability Diagnostic
arXiv:2610.04592v1 Announce Type: new Abstract: KRATOS is a group-structured bibliometric diagnostic that compares the distribution of documents (participation) with the distribution of citation weight (recognition) over…
- arXiv cs.AIresearchLatentIndex: Cross-Layer Sharing with Layer-Specific Selection for Sparse Attention
arXiv:2610.04635v1 Announce Type: new Abstract: Sparse attention reduces core-attention computation, but its indexers still incur repeated selection work and per-layer key-cache storage. Reusing selected indices across l…
- arXiv cs.AIresearchSEIS: Self-Evolving Inference Systems
arXiv:2610.04646v1 Announce Type: new Abstract: Inference systems determine how fast and how cheaply language models can be served, so making them faster has direct practical value. However, prior work focuses mostly on…
- arXiv cs.AIresearchPermVLA: Factorization Order as a Regularizer for VLA Learning
arXiv:2610.04659v1 Announce Type: new Abstract: Vision-language-action (VLA) policies commonly learn action chunks through a fixed left-to-right (LTR) factorization, although the same expert trajectory distribution admit…
- arXiv cs.AIresearchEfficient Neural Surrogates for Linear Radiation Transport on the Lattice and Hohlraum benchmarks
arXiv:2610.04665v1 Announce Type: new Abstract: Linear radiation transport equations (RTEs) form the simulation foundations underpinning design and analysis tasks in nuclear engineering, inertial confinement fusion, medi…
- arXiv cs.AIresearchA Tropical Geometry View of Forgetting: A Per-Unit Projector for Knowledge-Preserving Fine-Tuning
arXiv:2610.04670v1 Announce Type: new Abstract: Fine-tuning a language model on new text degrades what it already does. Replay-free projectors such as Adam-NSCL and GPM forbid one shared subspace of a layer's inputs in e…