The Validity Gap in Health AI Evaluation: A Cross-Sectional Analysis of Benchmark Composition
arXiv:2603.18294v1 Announce Type: new Abstract: Background: Clinical trials rely on transparent inclusion criteria to ensure generalizability. In contrast, benchmarks validating health-related large language models (LLMs) rarely characterize the "patient" or "query" populations they contain. Without defined composition, aggregate performance metrics...
Can LLM generate interesting mathematical research problems?
arXiv:2603.18813v1 Announce Type: new Abstract: This paper is the second one in a series of work on the mathematical creativity of LLM. In the first paper, the authors proposed three criteria for evaluating the mathematical creativity of LLM and constructed...
Agentic Framework for Political Biography Extraction
arXiv:2603.18010v1 Announce Type: new Abstract: The production of large-scale political datasets typically demands extracting structured facts from vast piles of unstructured documents or web sources, a task that traditionally relies on expensive human experts and remains prohibitively difficult to automate...
Balanced Thinking: Improving Chain of Thought Training in Vision Language Models
arXiv:2603.18656v1 Announce Type: new Abstract: Multimodal reasoning in vision-language models (VLMs) typically relies on a two-stage process: supervised fine-tuning (SFT) and reinforcement learning (RL). In standard SFT, all tokens contribute equally to the loss, even though reasoning data are inherently...
Interpretability without actionability: mechanistic methods cannot correct language model errors despite near-perfect internal representations
arXiv:2603.18353v1 Announce Type: new Abstract: Language models encode task-relevant knowledge in internal representations that far exceeds their output performance, but whether mechanistic interpretability methods can bridge this knowledge-action gap has not been systematically tested. We compared four mechanistic interpretability methods...
Continually self-improving AI
arXiv:2603.18073v1 Announce Type: new Abstract: Modern language model-based AI systems are remarkably powerful, yet their capabilities remain fundamentally capped by their human creators in three key ways. First, although a model's weights can be updated via fine-tuning, acquiring new knowledge...
Retrieval-Augmented LLM Agents: Learning to Learn from Experience
arXiv:2603.18272v1 Announce Type: new Abstract: While large language models (LLMs) have advanced the development of general-purpose agents, achieving robust generalization to unseen tasks remains a significant challenge. Current approaches typically rely on either fine-tuning or training-free memory-augmented generation using retrieved...
Consumer-to-Clinical Language Shifts in Ambient AI Draft Notes and Clinician-Finalized Documentation: A Multi-level Analysis
arXiv:2603.18327v1 Announce Type: new Abstract: Ambient AI generates draft clinical notes from patient-clinician conversations, often using lay or consumer-oriented phrasing to support patient understanding instead of standardized clinical terminology. How clinicians revise these drafts for professional documentation conventions remains unclear....
How Psychological Learning Paradigms Shaped and Constrained Artificial Intelligence
arXiv:2603.18203v1 Announce Type: new Abstract: The dominant paradigms of artificial intelligence were shaped by learning theories from psychology: behaviorism inspired reinforcement learning, cognitivism gave rise to deep learning and memory-augmented architectures, and constructivism influenced curriculum learning and compositional approaches. This...
Learning to Self-Evolve
arXiv:2603.18620v1 Announce Type: new Abstract: We introduce Learning to Self-Evolve (LSE), a reinforcement learning framework that trains large language models (LLMs) to improve their own contexts at test time. We situate LSE in the setting of test-time self-evolution, where a...
Frayed RoPE and Long Inputs: A Geometric Perspective
arXiv:2603.18017v1 Announce Type: new Abstract: Rotary Positional Embedding (RoPE) is a widely adopted technique for encoding position in language models, which, while effective, causes performance breakdown when input length exceeds training length. Prior analyses assert (rightly) that long inputs cause...
Engineering Verifiable Modularity in Transformers via Per-Layer Supervision
arXiv:2603.18029v1 Announce Type: new Abstract: Transformers resist surgical control. Ablating an attention head identified as critical for capitalization produces minimal behavioral change because distributed redundancy compensates for damage. This Hydra effect renders interpretability illusory: we may identify components through correlation,...
VC-Soup: Value-Consistency Guided Multi-Value Alignment for Large Language Models
arXiv:2603.18113v1 Announce Type: new Abstract: As large language models (LLMs) increasingly shape content generation, interaction, and decision-making across the Web, aligning them with human values has become a central objective in trustworthy AI. This challenge becomes even more pronounced when...
Conflict-Free Policy Languages for Probabilistic ML Predicates: A Framework and Case Study with the Semantic Router DSL
arXiv:2603.18174v1 Announce Type: new Abstract: Conflict detection in policy languages is a solved problem -- as long as every rule condition is a crisp Boolean predicate. BDDs, SMT solvers, and NetKAT all exploit that assumption. But a growing class of...
MolRGen: A Training and Evaluation Setting for De Novo Molecular Generation with Reasonning Models
arXiv:2603.18256v1 Announce Type: new Abstract: Recent advances in reasoning-based large language models (LLMs) have demonstrated substantial improvements in complex problem-solving tasks. Motivated by these advances, several works have explored the application of reasoning LLMs to drug discovery and molecular design....
Enactor: From Traffic Simulators to Surrogate World Models
arXiv:2603.18266v1 Announce Type: new Abstract: Traffic microsimulators are widely used to evaluate road network performance under various ``what-if" conditions. However, the behavior models controlling the actions of the actors are overly simplistic and fails to capture realistic actor-actor interactions. Deep...
Detection Is Cheap, Routing Is Learned: Why Refusal-Based Alignment Evaluation Fails
arXiv:2603.18280v1 Announce Type: new Abstract: Current alignment evaluation mostly measures whether models encode dangerous concepts and whether they refuse harmful requests. Both miss the layer where alignment often operates: routing from concept detection to behavioral policy. We study political censorship...
Path-Constrained Mixture-of-Experts
arXiv:2603.18297v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) architectures enable efficient scaling by activating only a subset of parameters for each input. However, conventional MoE routing selects each layer's experts independently, creating N^L possible expert paths -- for N experts...
MLOW: Interpretable Low-Rank Frequency Magnitude Decomposition of Multiple Effects for Time Series Forecasting
arXiv:2603.18432v1 Announce Type: new Abstract: Separating multiple effects in time series is fundamental yet challenging for time-series forecasting (TSF). However, existing TSF models cannot effectively learn interpretable multi-effect decomposition by their smoothing-based temporal techniques. Here, a new interpretable frequency-based decomposition...
AIMER: Calibration-Free Task-Agnostic MoE Pruning
arXiv:2603.18492v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) language models increase parameter capacity without proportional per-token compute, but the deployment still requires storing all experts, making expert pruning important for reducing memory and serving overhead. Existing task-agnostic expert pruning methods are...
Balancing the Reasoning Load: Difficulty-Differentiated Policy Optimization with Length Redistribution for Efficient and Robust Reinforcement Learning
arXiv:2603.18533v1 Announce Type: new Abstract: Large Reasoning Models (LRMs) have shown exceptional reasoning capabilities, but they also suffer from the issue of overthinking, often generating excessively long and redundant answers. For problems that exceed the model's capabilities, LRMs tend to...
Beyond Passive Aggregation: Active Auditing and Topology-Aware Defense in Decentralized Federated Learning
arXiv:2603.18538v1 Announce Type: new Abstract: Decentralized Federated Learning (DFL) remains highly vulnerable to adaptive backdoor attacks designed to bypass traditional passive defense metrics. To address this limitation, we shift the defensive paradigm toward a novel active, interventional auditing framework. First,...
Birthright citizenship: why the text, history, and structure of a landmark 1952 statute doom Trump’s executive order
Brothers in Law is a recurring series by brothers Akhil and Vikram Amar, with special emphasis on measuring what the Supreme Court says against what the Constitution itself says. For more content from […]The postBirthright citizenship: why the text, history,...
Justices to consider rules pardoning omissions by bankrupt debtors
Next week’s argument in Keathley v. Buddy Ayers Construction involves a technical question about bankruptcy procedure – the standards for overlooking the failure of a debtor in bankruptcy to mention […]The postJustices to consider rules pardoning omissions by bankrupt debtorsappeared...
Justices to consider the rights of asylum seekers at the U.S.-Mexico border
The Supreme Court will hear oral arguments next week in a challenge to the government’s policy of systematically turning back asylum seekers before they can reach the U.S. border with […]The postJustices to consider the rights of asylum seekers at...
Uninjured class members, hindsight harmlessness, presidential cronies, and the mistaken use of deadly force
The Relist Watch column examines cert petitions that the Supreme Court has “relisted” for its upcoming conference. A short explanation of relists is available here. There are 261 petitions and applications […]The postUninjured class members, hindsight harmlessness, presidential cronies, and...
SCOTUStoday for Thursday, March 19
Chief Justice Earl Warren was born on this day in 1891. Before Warren joined the court, he served as governor of California for a decade.The postSCOTUStoday for Thursday, March 19appeared first onSCOTUSblog.
Volume 2026, No. 1 – Wisconsin Law Review – UW–Madison
Contract Law and Civil Justice in Local Courts by Cathy Hwang & Justin Weinstein-Tull; Preempting Drug Price Reform by Shweta Kumar; Lessons Learned? COVID’s Continued Impact on Remote Work Disability Accommodations by D’Andra Millsap Shu; Unbundling AI Openness by Parth...
Whatever Did Happen to the Antitrust Movement?
ARTICLE Whatever Did Happen to the Antitrust Movement? Herbert Hovenkamp* Antitrust in the United States today is caught between its pursuit of technical rules designed to define and implement defensible economic goals, and increasingly political calls for a new antitrust...
The Singular Role of Public Pension Funds in Corporate Governance
Introduction U.S. public pension funds manage more than $6 trillion in assets.[1] The law, policy, and public debates about how they should manage this money are based on a theoretical model that is descriptively inaccurate and yields policy proposals that...