AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchOn the Privacy of LLMs: An Ablation Study
arXiv:2605.02255v2 Announce Type: replace-cross Abstract: Large language models (LLMs) face diverse privacy threats, yet existing studies typically evaluate individual attacks in isolation, offering limited guidance on w…
- arXiv cs.AIresearchCRAFT: Forgetting-Aware Intervention-Based Adaptation for Continual Learning
arXiv:2605.05732v3 Announce Type: replace-cross Abstract: Large language models (LLMs) can learn new tasks through fine-tuning, but learning tasks sequentially often leads to catastrophic forgetting of earlier ones. This…
- arXiv cs.AIresearchNeural Visual Decoding via Cognitive guided Adaptive Blurring and Information Constrained Alignment
arXiv:2605.16418v2 Announce Type: replace-cross Abstract: EEG-based visual decoding aims to establish a mapping between neural signals and visual semantics. However, it remains constrained by the dual challenges of sever…
- arXiv cs.AIresearchChaosBench-Logic v2: Evaluating LLM Logical Reasoning over Dynamical Systems at Scale
arXiv:2605.24305v2 Announce Type: replace-cross Abstract: Standard accuracy on binary reasoning benchmarks hides critical failure modes: prior collapse, inconsistency under paraphrase, and inability to reason about param…
- arXiv cs.AIresearchLLMSurgeon: Diagnosing Data Mixture of Large Language Models
arXiv:2605.30348v2 Announce Type: replace-cross Abstract: The pretraining data mixture of Large Language Models (LLMs) constitutes their "digital DNA", shaping model behaviors, capabilities, and failure modes. Yet this c…
- arXiv cs.AIresearchMoment-Video: Diagnosing Temporal Fidelity of Video MLLMs on Momentary Visual Events
arXiv:2606.02522v2 Announce Type: replace-cross Abstract: Video multimodal large language models (MLLMs) have made rapid progress in general and long-form video understanding, yet their ability to capture and use brief,…
- arXiv cs.AIresearchRethinking Calibration for High-Sparsity LLM Pruning: From Single Sources to Capability Coverage
arXiv:2606.03328v4 Announce Type: replace-cross Abstract: Post-training LLM pruning relies on a small calibration set, usually chosen heuristically and judged by WikiText perplexity or by mean accuracy on general-domain…
- arXiv cs.AIresearchANEForge: Python for direct computation on the Apple Neural Engine
arXiv:2606.17090v2 Announce Type: replace-cross Abstract: ANEForge is a Python package that programs the Apple Neural Engine (ANE), the fixed-function neural accelerator on every recent Apple device, directly and without…
- arXiv cs.AIresearchHitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift
arXiv:2606.25152v2 Announce Type: replace-cross Abstract: Deployed AI text detection approaches are supervised, relying on training-time access to labeled datasets of both human-written and AI-generated text. This approa…
- arXiv cs.AIresearchPlay2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?
arXiv:2606.26428v4 Announce Type: replace-cross Abstract: Multi-fingered robots promise the speed and dexterity of human hands, yet challenging problems such as precise assembly have remained out of reach. These tasks ar…
- arXiv cs.AIresearchFrom Signals to Transfer: A Factorised Study of Probe-Based Uncertainty Estimation in Large Language Models
arXiv:2606.27679v2 Announce Type: replace-cross Abstract: Probe-based uncertainty estimation (UE) has emerged as a prominent approach to detect hallucinations in Large Language Models (LLMs) by learning uncertainty from…
- arXiv cs.AIresearchSymbolic Mechanistic Data Attribution: Tracing Training Influence to Learned Behavioral Policies
arXiv:2606.29171v2 Announce Type: replace-cross Abstract: While existing data attribution methods can identify which training examples build specific mechanistic circuits, they cannot explain how training data shapes the…
- arXiv cs.AIresearchStabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping
arXiv:2606.29983v2 Announce Type: replace-cross Abstract: Looped Transformers, which repeatedly apply a shared transformer block, are an architecturally natural fit for variable-length algorithmic tasks. Although they ca…
- arXiv cs.AIresearchWords Speak Louder Than Code: Investigating Cognitive Heuristics in LLM-Based Code Vulnerability Detection
arXiv:2606.30587v2 Announce Type: replace-cross Abstract: Researchers and practitioners increasingly apply Large Language Models (LLMs) for automated vulnerability detection. Recent work has shown that LLMs are susceptib…
- arXiv cs.AIresearchLCPNet: Latent Consistent Proximal Unfolding Network for Infrared Small Target Detection
arXiv:2607.04603v3 Announce Type: replace-cross Abstract: Infrared small target detection (IRSTD) aims to identify long distance small targets from complex infrared backgrounds, and is a fundamental task in remote sensin…
- arXiv cs.AIresearchAirPASS: Over-the-Air Federated Learning via Pinching Antenna Systems
arXiv:2607.06768v3 Announce Type: replace-cross Abstract: This paper investigates over-the-air federated learning (AirFL) in wireless systems where the access point is equipped with a multi-waveguide pinching antenna sys…
- arXiv cs.AIresearchRight Answer, Wrong Tokens: Tail-Aware Credit Calibration for RLVR
arXiv:2607.07976v2 Announce Type: replace-cross Abstract: Although reinforcement learning (RL) improves the reasoning capabilities of large language models (LLMs), GRPO-style methods can suffer from a positive-credit con…
- arXiv cs.AIresearchBeatEdit: Symbolic Music Generation as Explicit Editing
arXiv:2607.11124v3 Announce Type: replace-cross Abstract: Music creation is fundamentally a process of revision. Yet symbolic music generation remains dominated by paradigms that produce complete sequences from scratch,…
- arXiv cs.AIresearchAmplitude-Only FFN Intervention for Tool-Structured LLM Inference: Method, Gated Evaluation Protocol, and Cross-Model Empirical Results
arXiv:2607.11183v5 Announce Type: replace-cross Abstract: We study amplitude-only rescaling of feed-forward network (FFN) activations to repair tool calls and other structured outputs, using multiplicative coordinate mas…
- arXiv cs.AIresearchCircuit Claims Depend on What Is Extracted and How It Is Compared
arXiv:2607.18921v3 Announce Type: replace-cross Abstract: Circuit extraction identifies model components whose presence preserves a target behavior under ablation, but preserving behavior does not single out one circuit.…
- arXiv cs.AIresearchVisual Contrastive Self-Distillation
arXiv:2607.21556v2 Announce Type: replace-cross Abstract: On-policy self-distillation (OPSD) removes the external teacher required by on-policy distillation (OPD), but still needs asymmetric information between teacher a…
- arXiv cs.AIresearchFilling Before Advancing: Capability-Gap-Driven Post-Training for Scenario-Specialized Remote Sensing MLLMs
arXiv:2607.22205v3 Announce Type: replace-cross Abstract: Remote sensing multimodal large language models (RS-MLLMs) have advanced aerial-image understanding, but scenario specialization remains constrained by scarce hig…
- arXiv cs.AIresearchValueFormer: Stage-Aware Value Labels for Online Failure Detection and VLA Post-Training
arXiv:2608.02958v2 Announce Type: replace-cross Abstract: Generalist vision-language-action (VLA) policies now transfer to new scenes with little task-specific training, but they still fail silently: from the action stre…
- arXiv cs.AIresearchDataKernelBench: Can LLMs Optimize Database Queries on GPUs?
arXiv:2608.25061v3 Announce Type: replace-cross Abstract: GPUs increasingly accelerate database systems, but query-specific peak performance still often relies on hand-written kernels. Existing LLM kernel benchmarks focu…
- arXiv cs.AIresearchICON decomposition: Multivariate concept auditing reveals shortcuts that neural networks rely on
arXiv:2608.26083v5 Announce Type: replace-cross Abstract: Deep neural networks often exploit spurious associations, a failure known as shortcut learning. Before deployment, models should therefore be audited for reliance…
- arXiv cs.AIresearchINTERVenE: Temporal-Abstraction-Interval Based Transformers for Short-Horizon Medical Event Prediction
arXiv:2608.29901v2 Announce Type: replace-cross Abstract: Electronic Health Record (EHR) prediction models in the intensive care unit must learn from sparse and irregular measurements while preserving the clinical meanin…
- arXiv cs.AIresearchTail-Replay: Escaping the Curse of Linear Attention in Prefix Caching for Hybrid LLMs
arXiv:2608.30310v2 Announce Type: replace-cross Abstract: Hybrid large language models interleave full-attention layers with linear-attention layers to reduce the cost of long-context inference. This structure complicate…
- arXiv cs.AIresearchEfficiently Estimating Optimal Hyperparameter Scaling Laws through Power-Law Entropy Search
arXiv:2609.01431v3 Announce Type: replace-cross Abstract: Optimal hyperparameter scaling laws describe how the best hyperparameters for large language model (LLM) training change with model and data scale, enabling pract…
- arXiv cs.AIresearchLanguage Models Can Control Their Own Attention
arXiv:2609.02737v2 Announce Type: replace-cross Abstract: Language models spend most of their attention on a small fraction of context, yet they read the entire KV cache to find the few tokens that matter. If the user as…
- arXiv cs.AIresearchTraining-Free Task Vectors for LLM Behavioral Control
arXiv:2609.09054v3 Announce Type: replace-cross Abstract: Task vectors enable post-training model editing by identifying semantically meaningful directions in weight space, typically computed as the difference between a…
- LessWrongresearchBad Guy Tokens
Based on a recent paper , training models to give bad advice in a narrow domain doesn't make the model or "the persona" broadly evil in a robust way. It just makes the particular system prompt [1] tokens you fine-tuned w…
- LessWrongresearchExploring Intrapersonal Game Theory
Work done as part of the MATS 9.1 extension program, mentored by Richard Ngo. The math in this post was developed with substantial AI assistance. Details are at the end. 0: Introduction A previous post of mine presents a…
- Simon Willison's Weblogresearchsparqlite: SPARQL 1.1 inside SQLite, written in Rust
Research: sparqlite: SPARQL 1.1 inside SQLite, written in Rust This research task is really an experiment in running Claude Code for web in "Ultracode" mode, as described in Orchestrate subagents at scale with dynamic wo…
- Engadget AIresearchAn OpenAI research agent 'stole' test Ethereum to finish a coding task
An OpenAI research model hit a nebulous milestone today by pilfering test Ethereum tokens.
- OpenAI NewsresearchBringing GPT-Rosalind to academic research labs
Introducing the Rosalind Academic Plan, offering eligible academic labs sponsored access to GPT-Rosalind and tools to advance life sciences research.
- Google ResearchresearchBringing the magic: Google breaks open a bottleneck in quantum computing
In a paper published in Science, Google has shown that Magic State Cultivation works on Google Quantum AI hardware.
- LessWrongresearchPredictor and picker – convergent architectural features of intelligent agents
Both your brain and an LLM first evaluate options and then pick one. Could this be a universal formula for agency? Note: While this post was not written for the LessWrong audience, I believe it might be of value to peopl…
- LessWrongresearchLinkpost: Neural Networks Are Already Clean
This is an extremely lengthy post by Nick Cammarata. Author's summary on twitter: interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don'…
- Google DeepMindresearchAccelerating the search for antimicrobial peptides with Google Antigravity
Accelerating the search for novel antimicrobial peptides using Google Antigravity for Science.
- Microsoft ResearchresearchProof-based general-purpose computing with Magnetar
Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains,…
- LessWrongresearchGLM-5.3 Sometimes Considers How Train-Time Actions Shape Future Behavior
Authors: Axel Højmark*, Ezra Newman* (*Equal contribution) While reviewing transcripts from standard behavioral evaluations of GLM-5.3, we noticed that the model occasionally reasons, unprompted, about how its actions wi…
- NVIDIA BlogresearchNVIDIA Expands Advanced Technology Access for Europe’s Scientific Community Through Open Learning Resources, Collaborations and Compute
Much of today’s scientific progress begins with computation. But turning accelerated computing’s potential into scientific progress — modeling climate systems, designing molecules, simulating materials and more — depends…
- arXiv cs.AIresearchBeyond Equation Search: Conjecture-Guided Discovery with SCOPE and ClosureBench
arXiv:2610.18938v1 Announce Type: new Abstract: Most symbolic regression methods optimize candidate equations, but improving an equation's fit does not establish whether its underlying scientific conjecture should be ret…
- arXiv cs.AIresearchDiscovering Undisclosed Concepts in LoRA Adapters for Diffusion-Based AIGC Models
arXiv:2610.18990v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) has made diffusion model personalization affordable and widely shareable, but it also creates moderation risks when harmful or problematic conten…
- arXiv cs.AIresearchRLPE: Teaching LLMs to Prompt with Reinforcement Learning
arXiv:2610.19062v1 Announce Type: new Abstract: Prompting is a widely used method for adapting general-purpose LLMs to domain-specific tasks. Effective prompts can be tedious to create manually, so prompt writing is incr…
- arXiv cs.AIresearchGraph-Based Orchestration of Heterogeneous AI Models: A Control Node Approach to Composite Intelligence
arXiv:2610.19136v1 Announce Type: new Abstract: We present a graph-based framework for Composite AI systems that integrate multiple AI techniques to address complex problems. We describe an architecture using specialized…
- arXiv cs.AIresearchCompiling Expert Procedures into Verified Programs: When a Small Model Is Enough
arXiv:2610.19148v2 Announce Type: replace Abstract: In procedural tool-use domains, how much of the gap between base models is procedure rather than capability? We build an explicit program from a teacher's traces and ru…
- arXiv cs.AIresearchListQA: A Benchmark for Evaluating List-Formatted Factual Knowledge Retrieval in Large Language Models
arXiv:2610.19165v1 Announce Type: new Abstract: Many real-world questions demand not a single fact but an organized collection of them, yet existing factual knowledge benchmarks almost exclusively target single answer re…
- arXiv cs.AIresearchLearning from Self-Generated Responses: Reliability--Uncertainty Self-Distillation for Reasoning
arXiv:2610.19176v1 Announce Type: new Abstract: Reference solutions and question-specific rubrics are not always available to guide the teacher in self-distillation. A large language model's own solution attempts offer a…
- arXiv cs.AIresearchMulti-System Cognitive Steering for Generative Communication
arXiv:2610.19177v1 Announce Type: new Abstract: A fundamental bottleneck for translating thoughts to speech is the high physical cost of motor input forcing a strict tradeoff between communication speed and expressivity.…