A-Evolve-Training
“We report an autonomous system that runs this loop with no human in the loop, post-training a 30B Nemotron across four rounds over multiple weeks.”
The authors, in the abstract
Autonomous Post-Training of a 30B Model
AI Research Radar
June 24–30: 10 frontier papers out of 1,954 read.
Frontier
Papers rated as moving the field
How we chooseSWE-INTERACT
“Across a suite of frontier and open-weight models, we find that strong performance on single-turn SWE tasks does not reliably transfer to multi-turn, user-driven workflows: the best-performing models solve roughly 50% of single-turn baseline tasks but only 25% of the corresponding SWE-Interact tasks.”
X-Stream
“Our extensive online inference experiments reveal a stark reality: state-of-the-art MLLMs struggle significantly with concurrent streams, achieving only about 50% score and exhibiting poor proactive ability.”
Beyond Function Calling
“Experiments reveal a substantial reliability gap: agents that perform well with reliable tools often fail under recoverable hazards.”
Know2Guess
From Sounds to Scenes
“Experiments across multiple LALMs demonstrate that effective auditory scene understanding requires integration over all auditory layers, rather than reliance on speech or sound alone, underscoring the necessity of CASU for advancing complex audio understanding in LALMs.”
Qwen-AgentWorld
“Empirical results demonstrate that Qwen-AgentWorld significantly outperforms existing frontier models.”
Safety from Honesty in a Disinterested AI Predictor
On the Inseparability of Instructions and Data in Shared-Embedding Sequence Models
Inherited Circuits, Learned Semantics
The week by field
Jun 24–30: 152 cleared, 10 frontier
The week’s papers have not been sorted into fields yet.
Models under test
Papers this week that name the model
Labs publishing
Papers this week with an author at the lab
Also cleared
142 papers
- HarmVideoBench: Benchmarking Harmful Video Understanding in Large Multimodal ModelsByteDance, Tencent
- JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree DraftingStepFunCode
- Improved Large Language Diffusion ModelsByteDanceCode
- PolicyAlign: Direct Policy-Based Safety Alignment for Large Language ModelsAlibabaCode
- Building to the Test: Coding Agents Deliver What You Check, Not What You RequestedMicrosoft
- Experience Graphs: The Data Foundation for Self-Improving AgentsMeta
- Yuvion VL: A Multimodal Foundation Model for Adversarial Content and AI SafetyAlibaba
- Support-Constrained RL Enables Real-World Policy Improvement without Real-World ExperienceAmazon
- End-to-End Dynamic Sparsity for Resource-Adaptive LLM InferenceMeta
- Discovering Millions of Interpretable Features with Sparse AutoencodersAlibaba
- Scalable Behavior Cloning with Open Data, Training, and EvaluationAmazon
- PhoneBuddy: Training Open Models for Agentic Phone UseTencent
All 142 papers
- Power-Flexible AI Data Centers: A New Paradigm for Grid-Responsive ComputeNVIDIA
- Quantization Inflates Reasoning: Token Inflation as a Hidden Cost of Low-Bit Reasoning ModelsMicrosoft
- Tracing Target Answers in Poisoned Retrieval Corpora via Token Influence AttributionIBM
- Autodata: An agentic data scientist to create high quality synthetic dataMeta
- Self-Recognition Finetuning can Prevent and Reverse Emergent MisalignmentGoogle
- Detecting Clinical Hallucinations in LVLMs via Counterfactual Visual Grounding UncertaintyCode
- Faults in Our Formal Benchmarking: Dataset Defects and Evaluation Failures in Lean Theorem ProvingCode
- TraceLab: Characterizing Coding Agent Workloads for LLM ServingCode
- CBD: API-Only LLM Black-Box Unlearning through Controlled Behavioral DivergenceCode
- Do Safety Guardrails Need to Reason? LeanGuard: A Fast and Light Approach for Robust ModerationCode
- Prompt Injection in Automated Résumé Screening with Large Language Models: Single and Multi-Injection SettingsCode
- SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic ReasoningCode
- Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution ShiftCode
- REALM: A Unified Red-Teaming Benchmark for Physical-World VLMsCode
- Governance Decay: How Context Compaction Silently Erases Safety Constraints in Long-Horizon LLM Agents
- Can LLMs Reliably Self-Report Adversarial Prefills, and How?
- It Lied to a Doctor to Buy Poison Ingredients: Quantifying Real-World Misuse of Phone-use Agents
- "AI Watermarking": Bridging Policy Discourse and Technical Capabilities
- When Medical Safety Alignment Fails: A Benchmark for Evaluating LLMs on High-Risk Medical Queries
- PIXELRAG: Web Screenshots Beat Text for Retrieval-Augmented Generation
- Auditing LLM-Governed Social Robots with Culture-Specific Moral Gradients
- Sifei at SemEval-2026 Task 8: Hybrid Retrieval and Query Rewriting for Multi-Turn RAG
- LLM based Knowledge Graph Approach to Automating Medical Device Regulatory Compliance
- Data Provenance for Image Auto-Regressive Generation
- Tool Use Enables Undetectable Steganography in Multi-Agent LLM Systems
- When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs
- Decomposing Memorization Reduction in Privacy-Preserving Fine-Tuning of SLMs for CSIRTs
- RIPA: Sensory-Vector Prompt Injection Attacks on LLM-Controlled ROS 2 Robots
- Why Trust Your Agent? Empirical Security Gains from TRiSM-Guided Agentic Workflows in Healthcare
- Constrained Tabular Diffusion for Finance
- Capability Gates Are Not Authorization: Confused-Deputy Failures in LLM Agent Frameworks
- An AI agent for treatment reasoning over a biomedical tool universe
- Agent Safety Is Action Alignment
- The Heterogeneous Safety Impacts of Benign Multilingual Fine-Tuning
- Defeat Devices in AI Systems
- Modification-Considering Value Learning for Reward Hacking Mitigation in RL
- FlipGuard: Defending Large Language Models Against Quantization-Conditioned Backdoor Attacks
- Can LLMs Hire Fairly? Racial Bias in Resume Screening
- When Stopping Fails: Rethinking Minimal Risk Conditions through Human-Interactive Autonomous Driving for Safe Transportation Systems
- Agent Security Meets Regulatory Reality -- A Practitioner Systematization of Autonomous-Agent Threats and Controls in Regulated Financial Systems
- Manufactured Confidence: How Memory Consolidation Turns Hearsay into Confident Facts
- Bit-ViP: Leveraging Bit-planes to Preserve Visual Privacy in Images through Obfuscation
- Closing the Activation-Cone Blind Spot: Response-Time Probing and Unified Defense
- Benchmark AUC Is Not Deployable Reliability: A Cross-Dataset Audit of Off-the-Shelf Features for Surveillance Video Anomaly Detection
- Proteus: Automated Adversarial Robustness Testing for Audio Deepfake Detectors
- Speculative Pre-Positioning: Decoding Stateful Sessions to the Next Decision Point Off the Critical Path
- SurrogateShield: Beyond Redaction for High-Utility, Privacy-Preserving LLM Interactions
- An Empirical Evaluation of Prompt Injection Vulnerabilities in Large Language Models Across Multilingual and Obfuscated Attack Scenarios
- Fuzzing Large Language Models to Elicit Hidden Behaviours
- Resolution Thresholds in VLM Detection of Harmful ASCII Art Across Construction Modes and Languages
- As We May Search
- CAREBench: A Child-Safety Risk Benchmark for Language Models
- Rethinking Generative Reconstruction Attacks against Graph Neural Network Models
- MemLeak: Diagnosing Information Leaks in Multimodal Agent Memory
- Concept Removal Guidance: Evidence-Calibrated Negative Guidance for Safe Diffusion Sampling
- Not-quite-human tastes: the stylized omnivorousness of LLM survey surrogates
- On the Internet, Nobody Knows You're an LLM Bot: Unmasking Web Agents with Multi-Layer Fingerprinting
- From Accuracy to Visual Dependence: Auditing and Filtering Modality Collapse in Traffic VideoQA
- Your Space is My Zone: Demystifying the Security Risks of AI-Powered Applications on Pre-Trained Model Hubs
- Arko-T: A Foundation Model for Text-to-Structured 3D Generation
- COHORT: Collaborative Orchestration for Hardening via Offensive Replay on Emulated Topologies
- Forensic Trajectory Signatures for Agent Memory Poisoning Detection
- Sequential Planning via Anchored Robotic Keypoints
- Artificial Intelligence Index Report 2026
- PYPILINE: Malicious PyPI Package Detection via Suspicious API Knowledge and Agent Workflow
- AI Snitches Get Glitches: Towards Evading Agentic Surveillance
- Helpfulness Hurts: Domain-Dependent Degradation of Mid-Trained Compassion Values Under Post-Training
- MIRAGE: Protecting against Malicious Image Editing via False Moderation
- Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation
- Cluster, Route, Escalate: Cascaded Framework for Cost-Aware LLM Serving
- TruEye: Fine-Grained Detection of AI-Generated Human Subjects in Images
- When the Aggregator Cheats: Data-Free Backdoors in Federated LLM-based QA Systems
- Retroactive Advantage Correction: Closed-Form V-Trace Bias Correction for Delay-Aware RLHF
- Qwen-Image-2.0-RL Technical Report
- Halt Fast! Early Stopping for Certified Robustness
- SpikeVLA: Vision-Language-Action Models with Spiking Neural Networks
- Agentic AI-Powered Re-Identification: An Emerging, Scalable Threat to Mobility Microdata Privacy
- Robust Harmful Features Under Jailbreak Attacks: Mechanistic Evidence from Attention Head Specialization in Large Language Models
- Agent-Native Immune System: Architecture, Taxonomy, and Engineering
- Benchmarking Open-Weight Foundation Models for Global AI Technical Governance
- Dream machine -- the next creative economy
- Divergent Recommendations, Convergent Diagnoses: Cross-Provider Failure-Mode Convergence in AI Commercial Recommendation
- The Governance Inversion Hypothesis: Why More AI Regulation May Produce Less Organisational Control
- Necessary but Not Sufficient: Temperature Control and Reproducibility in LLM-as-Judge Safety Evaluations
- TEMPO-Diffusion: Temporally Exposed Malicious Poisoning of Diffusion Models
- Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems
- Instruction Bleed: Cross-Module Interference in Prompt-Composed Agentic Systems
- When Agents Meet Electric Bus Fleet Operations: Pricing Behavior, Trade-offs, and Policy Implications in an Aggregator Framework
- Otter Weather: Skillful and Computationally Efficient Medium-Range Weather Forecasting
- VIGIL: Runtime Enforcement of Behavioral Specifications in AI Agent Skills
- Tactile-WAM: Touch-Aware World Action Model with Tactile Asymmetric Attention
- MIRROR: Novelty-Constrained Memory-Guided MCTS Red-Teaming for Agentic RAG
- The Capability Frontier: Benchmarks Miss 82% of Model Performance
- Confidence-Aware Tool Orchestration for Robust Video Understanding
- Chai: Agentic Discovery of Cryptographic Misuse Vulnerabilities
- Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactions
- ShareLock: A Stealthy Multi-Tool Threshold Poisoning Attack Against MCP
- Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)
- From Celebrities to Anyone: Characterizing AI Nudification Content, Technology, and Community Dynamics on 4chan
- Small edits, large models: How Wikipedia advocacy shapes LLM values
- Do Thinking Tokens Help with Safety?
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- Forget to Improve: On-Device LLM-Agent Continual Learning via Budget-Curated Memory
- Transferable Attack against Face Swapping in an Extended Space
- Generative AI for Safe and Photorealistic Drone Light Shows
- How Reliable Is Your Jailbreak Judge? Calibration and Adversarial Robustness of Automated ASR Scoring
- Concept Removal for Frontier Image Generative Models
- Leaking Circuit Secrets: Gradient Leakage Attacks on Graph Neural Networks
- SurgAtlas: A Large-Scale Surgical Video-Language Dataset with 2,391 Hours of Open and Minimally Invasive Surgery
- Privacy Vulnerabilities of Attention Layers in Tabular Foundation Models and Protection of High-Risk Queries
- Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning
- The Unfireable Safety Kernel: Execution-Time AI Alignment for AI Agents and Other Escapable AI Systems
- Real-Time Voice AI Hears but Does Not Listen
- Do Foundation Models See Biology? Evaluating Attention Coherence with Spatial Transcriptomics in Glioblastoma
- How Much Can We Trust LLM Search Agents? Measuring Endorsement Vulnerability to Web Content Manipulation
- One Year Later...The Harms Persist, But So Do We!
- RIFT-Bench: Dynamic Red-teaming For Agentic AI Systems
- Cyclic Denoising Reveals Ultrastable Memories in Diffusion Models
- AutoSpec: Safety Rule Evolution for LLM Agents via Inductive Logic Programming
- Securing LLM-Agent Long-Term Memory Against Poisoning: Non-Malleable, Origin-Bound Authority with Machine-Checked Guarantees
- When Helpfulness Overrides Causal Caution: Context-Dependent Suppression and Recovery in LLMs
- Age of LLM: A Strategic 1v1 Benchmark for Reasoning, Diplomacy and Reliability of Large Language Models under Fog of War
- Poisoned Playbooks: Demystifying Knowledge Poisoning Effects on AI Security Agents
- Detecting AI Coding Agents in Open Source: A Validated Multi-Method Census of 180 Million Repositories
- S1-Omni-Image: A Unified Model for Scientific Image Understanding, Generation, and Editing
- The African Language Tax: Quantifying the Cost, Latency, and Context Penalty of Tokenizing African Languages in Frontier LLMs
- Red-Teaming the Agentic Red-Team
- OpenThoughts-Agent: Data Recipes for Agentic Models
- BenchX: Benchmarking AI Models for Cancer Detection and Localization with Demographic and Protocol Biases
- InSight: Self-Guided Skill Acquisition via Steerable VLAs