Archive
Everything we've scored that's past the 48-hour window — open to everyone, no account needed. The live feed only shows what's moving right now; this is the record behind it.
- 19Stress-Testing Efficient Responsible-AI Evaluation: When Compute Savings Change Benchmark Conclusions
- 19Belief-Aware Multi-Agent Path Finding under Map Uncertainty
- 19Don't Be Nice
- 19Probe-Space Preconditioning for Fast and Stable Zero-Order Training
- 19Trace Integrity for LLM Data Agents: A Vision for Auditable Structured Reasoning in Real-World Systems
- 19Tail-Influence Sampling for CVaR Policy Evaluation
- 19WildMAE: Structure-Aware Masked Autoencoding for Wildlife Image Reconstruction
- 19NeuronEye: Query-Guided Visual Concept Activation for Vision-Language Reasoning
- 19Effective Dense Retrieval using Only In-Context Examples
- 19Bellman Calibration for Marginalized Importance Weighting in Offline Reinforcement Learning
- 19SEAR: Self-Evolving Agentic Robotics
- 19Improving Cross-Problem Vehicle Routing with Locally Augmented Preferences and Representation Disentanglement
- 19Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models
- 19nanoMuse: An Open-Source Personal Agent for Every Device You Own
- 19Aspire: Can Models Self-Evolve from Vague Goals?
- 19PaPi: Parallel Picard-Duhamel Integration for Fast Neuro-Spectral Architectures
- 19Explore Broadly, Reason Sharply: Push Small Models toward the Frontier via Sampling
- 19Coupled Calibration and Learning: Mitigating Teacher Bias in LLM Distillation without Target-Domain Reward Feedback
- 19Consilience for Verifier-Free Test-Time Scaling
- 19JustFit: 200K-Token LLM Serving on a 24 GiB Laptop with Just-in-Time State Management
- 19SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?
- 19What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models
- 19Harness Learning Enables Generalizable Test-Time Adaptation
- 19Pruning for Efficiency, Paying in Fairness: Demographic Disparities in Pruned Speech-LLMs
- 19Before You Poll with LLMs: A Deliberative Diagnostic Framework
- 19Correct Answers, Invalid Traces: What Verifiable Grade-School Math Reveals About Chain-of-Thought Traces
- 19Bias-Induced Crossover in Absolute Capacity of Dense Associative Memory
- 19MeClear: Cooperative Game-Theoretic Attribution and Risk-Aware Memory Clearance for Long-Horizon LLM Agents
- 19Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution
- 19cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents
- 19Agreement Is Not Validity: Cross-Model LLM Consensus in Diagnosing Student Failure Modes in K-12 Math Tutoring Dialogue
- 19When Does Scale-Invariant Optimization Become Unstable? An Exact Schedule Law with Weight Decay
- 19PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents
- 19Linguistic Loopholes in LLM Unlearning: From a 174-Language Benchmark to Coverage-Aware Unlearning
- 19PMosFM: Preconditioned Manifold Matching for One-Step Physics-Constrained Generation
- 19How Local Mixing Encodes Relative Position in Global NoPE Attention
- 19Fairness in Link Prediction Beyond Demographic Parity: A Reproducibility Study
- 19rustgrep - structural grep for Rust source
- 19Decoding-Level Taboo: A Diagnostic Stress Test for LLM Robustness
- 19RLLBC-Lib: An Educational Code Library for Reinforcement Learning and Learning-Based Control
- 19Frozen Flows Forget: Diagnosing and Restoring Lost Motion in a Latent-flow World Model
- 19Shockingly Simple Self-retrospection Improves Agentic Models Without RL
- 19Parameterized Complexity of $L_p$-Lipschitz Constants for Input Convex Neural Networks and $L_p$-Norm Maximization over Zonotopes
- 19zLend: A Dual-Scope Cash-Flow Reconstruction Framework for On-Chain Credit Underwriting
- 19CAS II: Symmetric Partitions as Kolmogorov Models
- 19What Do RTL Repair Benchmarks Actually Measure?
- 19FinAutoRubric: Expert-Guided Automatic Rubric Generation for Evaluating Financial Research Agents
- 19DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination
- 19Copy the Same, Distill the Difference: Initializing Linear Vision Transformers
- 19From Routing Signals to Selective Review: Visual regrounding in MoE VLMs
- 19DSLE: A Learning Environment for Dark Souls Boss Encounters
- 19Disentangling Computation in Multi-Task Neural Networks with the Green's Operator
- 19ReCIRC: Rectified Conformal Risk Control
- 19Agent-Editing World Model: Rethinking World Modeling for LLM Agents
- 19How Much Is an AI Token Worth? Scaling Laws for Wild AI-Generated Web Text
- 19Neural Implicit Energy Optimization for LiDAR Mapping
- 19Double descent is the principle of least action
- 19Registration is a Representation Problem: Solving Whole-Body PET/CT Registration
- 19Probabilistic Linear Explanations
- 19The Structure of Quantization Damage in LLMs: Why the Next Bit Should Be Spent Globally
- 19Robust CurveMoE: Multi-Norm Adversarial Defense for Mixture-of-Experts Models via Mode Connectivity
- 19Evolving Agent Interfaces
- 19Turning TS2Vec into a Foundation Model for Time Series Classification
- 19Improving Test-Time Scaling with Adaptive Looped Transformers
- 19Det-LIME: Detector-Aware, Multi-Instance Local Interpretable Model-Agnostic Explanations for Automated Marine Mammal Detection
- 19Stochastic World Models for Verifying Vision-Based Neural Feedback Systems
- 19How Much of a Harness Does a Strong Agent Need for Autonomous ML Engineering?
- 19SkillScaffold: Scaffolding Agentic Reinforcement Learning with Policy-Adaptive Skill Credit Assignment and Selective Injection
- 19Designing Proactive Thought Partners for Writing
- 19Optimizing Team Behavior via Extremum-seeking Control in Multi-agent Reinforcement Learning
- 19Towards Communication-Efficient Social Intelligence in Language Agents
- 19WUSH-KV: KV Cache Quantization with Data-Adaptive Transforms
- 19Depth Memory
- 19KV-streams for Efficient Compaction in Agentic Reinforcement Learning
- 19Canonical Color as a Lens into Concept Decodability in Vision Encoders and VLMs
- 19Looped Diffusion Transformer
- 19DynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agents
- 19StudentSim: Training LLM-based Student Simulators
- 19Value-Rich, Differently Organized: Measuring Value Expression and Organization in AI-Generated Debates
- 19How to Loop MoE: Flatten the Experts, Untie the Attention
- 19Neural Harmonic Measure Operator
- 19Class Activation Mapping in Explainable Computer Vision: A Method-Centered Review of CNN, Transformer, and Foundation-Model-Era Visual Explanations
- 19Learning to Coach for Experiential Learning
- 19A Generalization of Amari's Bayesian Duality
- 19Statistical Vignettes on the Grokking Phenomenon
- 19Compression Footprints as Security Signals for Model-Poisoning Defense in Federated Learning
- 19Statistical Learning of Contractive Dynamical Representations for Composite Adaptive Control
- 19Scaling Long-Form Story Generation via Narrative State Tracking
- 19TokenCast: Forecasting Token Consumption During LLM Agent Execution
- 19Benchmarking Arabic Dialects for Math Problems
- 19Unifying Distributional Training for One-Step Visual Generation
- 19A Framework for Designing Reward Functions: From Objectives to Features to Human-Aligned Reward Functions
- 19SPO++: Stream-Aligned Policy Optimization for Asynchronous Agentic RL
- 19Retrieving Biblical Intertextual References in Karen Blixen's Seven Gothic Tales
- 19Scaling Laws for Looped Mixture of Experts
- 19Self-Evolving Agent Harness
- 19Learning Native Reflection in Unified Models with Interleaved Reinforcement Learning
- 19EIP-8379: Top-up Sync
- 19PDMD: Projected Distribution Matching Distillation for Video Diffusion Models
- 19The Value of Human Expertise
- 19Telescopic Language Models
- 19FurE: Efficient Instance-Specific 3D Fur Reconstruction without Animal-Fur Datasets
- 19How Much Rank Does LoRA Need? Rank-Error Bounds for Transformer Attention
- 19Achieving an $O(1/N)$ Optimality Gap in Average-Reward Weakly-Coupled MDPs
- 19$R^3$: Training Robots to Reason in Natural Language via Reinforcement Learning
- 19The canonical facets of multi-separator polytopes
- 19Prediction-powered inference for time series across space
- 19Mechanism Design for Alignment and Control
- 19Multi-Agent Flow Matching with Decoupled Generative Guidance
- 19Nearly Tight Rademacher Bounds for Sparsely Activated Neural Networks
- 19MatLoom: Layered Text-to-Material Generation in a Compact Program Space
- 19K-Bench: a clinically calibrated benchmark for evaluating large language models in high-risk mental health conversations
- 19Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation
- 19Bridging the Gap Between Homogeneous and Heterogeneous Asynchronous Optimization Is Surprisingly Difficult
- 19The Rise of Verbal Reinforcement Learning
- 19Cogentic: Multi-Agent Orchestration for Automated Proof Discovery
- 19Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs
- 19ExecCritic: Learn to Test, Test to Improve for Coding Agents
- 19WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents
- 19Looped Uniform Diffusion Language Models
- 19GoiBERT: Compact Vietnamese Language Models via Two-Stage Knowledge Distillation
- 19Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails
- 19Disentangling Paradigm, Identifier, and Decoding in Generative Retrieval
- 19LongHarness Bench: Stress-Testing Language Model Harnesses for Long-Context Reasoning
- 19Task-Directed Residual AddUNet:Perfect-Reconstruction Routing for Full-Rate Representations
- 19Breaking the Uniformity Trap: Scaling Video Diffusion Model via SplitMoE
- 19Entropy-Regularized Rank-Masked Policy Optimization for Test-Time Reinforcement Learning in Code Generation
- 19Turbo Harness: Instance-Adaptive Harness Optimization
- 19Loopholing All the Way: Language Modeling with pure Self-Conditioning
- 19CordisBench: Can Language Models Reason About Component Lifecycles in Dynamic Agent Harnesses?
- 19Read the Room Before Your Model Generates
- 19AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation
- 19A General Kernel Framework for Non-CND Distance Measures Using |D|-Dimensional Sparse Landmark Embeddings
- 19Non-Crossing Deep Quantile Regression for Distributional Survival Prediction
- 19Co-Evolving Paths and Flows via Path-Flow Alignment
- 19TimescaleDB 2.27 Added Bloom Filters to UPDATE and DELETE. Your EXPLAIN Won't Tell You If They Work Unless You Know These Counters.
- 19Adaptive Critical Token-Aware Retrieval for Repository-Level Code Generation
- 19Learning Meta-Skills for Agent Harness Design in Test-Time AI4AI
- 19Efficient SWE Agent Benchmarking via Trajectory-Aware Evaluation
- 19Semantic Observability for Autonomous Systems Diagnostics
- 19Beyond Scores: Understanding LLM-as-a-Judge Mechanisms in Summarization Evaluation
- 19LongAgent: History-Guided Agentic Search for Longitudinal Outcome Prediction
- 19A Data-Driven Framework for Identifying and Prioritizing RPA Opportunities in Healthcare Processes
- 19"Train classical, deploy quantum" requires rethinking generalization
- 19SWE-Serve: Benchmarking Agentic Engineering For Production Inference Serving
- 19Joint Neural Posterior Estimation and Model Selection for Chemical Reaction Networks
- 19Is Weight Tying Still Beneficial for Decoder-Only LLMs in Private Settings Under DP-SGD?
- 19When Forgetting is not Catastrophic: On the Mechanics of Spurious Forgetting
- 19Charts as Epistemic Actions for Experience-Driven Agentic Data Analysis
- 19Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses
- 19NOAH: Learning the Full Patient Journey. A Longitudinal Multimodal Time-Aware Model for Representation and Forecasting
- 19Holdout Best-of-N: Unbiased Evaluation and Its Cost
- 19Thinking Before Thinking: Scaling Agentic Inference Through Meta-Reasoning
- 19Constructing Dynamic Master Logic Models as Knowledge Graphs for Complex System Diagnostics Using Retrieval-Augmented Large Language Models
- 19Parallel AI Agent Collaboration as Graph-Structured Distributed Block Optimization
- 19Pretraining Latent Information Feedback Transformers with Teacher Supervision
- 19WorldSolver: Can LLM Agents Simulate the Physical Dynamics via Solver Generation?
- 19EvoDuet: Bilevel Co-Evolution of Web Searching and Task Solving for Scientific Discovery
- 19CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents
- 19SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue
- 19Repairability of Inexact Solvers in Recursive State Estimation with Machine Learning
- 19From Classifier Heads to Encoder Prototypes: Personalized Federated Learning with Head-Derived Clustering
- 19Circuits of LLM Classifiers
- 19MUSE: Benchmarking Large Vision-Language Models on Multi-Modal Understanding in Situated Education
- 19Agensh: Scaling Organizational Intelligence to 1,024 Agents
- 19GENCO - A Unified Neural Solver Embedded in a Development Framework for Steady-State Grid Analysis
- 19Does an Agent's History Tell You When Compaction Will Hurt? A Modest, Bounded Effect on the TRACE Paired-Replay Corpus
- 19NarraSpace: Multimodal Scene-Grounded Evidence for Picture-Description Dementia Classification
- 19Image Classifiers are Efficient Self-Supervised Video Representation Learners
- 19Fine-Tuning Whisper for Automatic Speech Recognition in Baniwa: A Preliminary Study
- 19A Decentralized Partially Observable Team Decision Methodology with Delayed Information Sharing
- 19Efficient Edge-aware Attention Network for Graph Generation
- 19OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding
- 19ViTeX-Bench: Benchmarking High-Fidelity Video Scene Text Editing
- 19Securing quantum error correction against misleading advice from AI agents
- 19Studying Image Tokenizers as Visual Languages in Unified Multimodal Models
- 19Context-Continuous Preference Learning for Exoskeleton Personalization
- 19Beyond the Timeline: Augmenting Long-Video Memory with Grounded Entity Biographies
- 19Adaptive Multi-Token Prediction: How far should we go?
- 19EmoRES-TTS: Residual-Enhanced Vector Steering for Emotional Speech Generation
- 19Removing Timing Shortcuts Improves Non-Invasive Brain-to-Text
- 19Towards Computational Provenance: Carrying Causal-State Evidence in Generated Text
- 19Ranking-Aware Prompt Optimization for Multimodal Clinical Diagnosis
- 19Semifactual Credit-Augmented Policy Optimization
- 19A Spectral Theory of Distortion in LLM Graph Reconstruction: Sharp Bounds and Empirical Characterization
- 19What FID Hides: Detecting, Ranking, and Diagnosing Deviations in Generative Evaluation
- 19Cropland PAtteRNS: Parallel Dimensional Attention Networks and Attention to Dataset Disparity for Crop Segmentation in Satellite Imagery Time Series Data
- 19From Values to Benchmarks: Evaluating Large Language Models for Governmental Use in Dutch
- 19RoboSimStudio: An Interactive Toolkit for Rapid Real2Sim of Robot Manipulation Tasks
- 19LeapQuant: Efficient Linear Attention with Accurate Recurrent State Quantization
- 19STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization
- 19Breakdown of Local Denoising as Semantic Speciation
- 19Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering
- 19Multimodal Model Diffing for Feature Discovery and Control
- 19Skill-Space Shooting for Autonomous Robot Policy Improvement
- 19Redistribution-based Cost Inference Improves Sparse Safe Offline RL
- 19AI4AI at Test-Time: Strong-to-Weak Capability Transfer via Harnesses
- 19Beyond Naturalness: Probing Automated Text-To-Speech Evaluators on Linguistically Grounded Dimensions
- 19When Does Bigger Help? A Controlled Study of LLM Scale for Ontology Learning
- 19LimiX-2: A Contextual Mechanism Network Towards General Structured-Data Intelligence