h1: A Taste of the Bitter Lesson in Automated Theorem Proving
While progress in informal mathematics has come largely from scaling the chain-of-thought and agentic reasoning of frontier language models, formal theorem proving has increasingly...
Papers
Memory Tax: Dissecting the Net Value of Agent Memory
Memory is increasingly used in long-horizon agents, yet the net value of building memory over direct retrieval from raw history remains largely unclear. Through a large-scale evalu...
HuGe: Scaling Humanoid Generalist Vision-Language-Action Models with Egocentric Experience
Extending Vision-Language-Action models (VLAs) to humanoid robots requires coordinated whole-body control and diverse demonstrations that are costly to collect in real-world scenar...
Identically Trained Language Models Break on Different Jailbreaks
Jailbreak benchmarks test one checkpoint and treat which jailbreaks break it as a property of its pretraining data and safety training. We ask whether another pretraining run of th...
InteractiveWorld: A Verifiable Human-Centric Environment for Building Better Interactive MLLM Assistants
People have long dreamed of an assistant that is always there to guide them and remind what they forget, from new dishes to lost keys. Powered by multimodal large language models (...
Continual Learning Exam: Benchmarking Long-Horizon RSI and Experiential Learning in LLM Agents
Continual learning is the ability of LLM agents to improve over time by acquiring, refining, and applying knowledge from experience while retaining useful prior knowledge. It is es...
ParetoRSI-Gym: Can AI Agents Improve Language Models Across Multiple Domains?
Recursive self-improvement (RSI) requires an agentic system to design, train, and evaluate its own successor. Data-driven post-training is a central part of that loop, yet existing...
Spatial Audio Generation and Editing with Executable Plans and Self-Distilling Transfusion
Spatial audio authoring coordinates source decisions with soundfield synthesis. For editing, plans specify what to change and retain, while reference audio supplies further acousti...
Humans in Space: Can MLLMs Construct Human-Centric Spatial Understanding via Active Exocentric Exploration?
While Multimodal Large Language Models (MLLMs) excel at zero-shot reasoning, their spatial capabilities are primarily evaluated through static, passive observation. In complex, hum...
Matched Capabilities, Different Behaviors: A Controlled Study of MCP and CLI Interfaces for Language Agents
Language agents increasingly act through external tools and interactive environments, making the interface between the model and its executable capabilities part of the decision pr...
Broken on Arrival: Non-Functional Quantized LLM Artifacts in Public Model Registries and How to Catch Them
Developers download quantized large language models (LLMs) from public registries to run them locally. None of the registries we examined documented functional testing of their fil...
LOLPlay: Benchmarking Real-Time Visual Agents in League of Legends
Visual agents need to coordinate reasoning and action while their environment evolves during inference. We introduce LOLPlay, a benchmark in native League of Legends with 36 champi...
AgentBaton: Efficient Model Collaboration in Agentic Tasks
Solving long-horizon agentic tasks requires language agents to interact with their environments over multiple steps, and reliable performance often comes from the expensive capable...
Beyond Benchmarks: Can Small Open-Weight Models Get Real Coding Work Done?
Making an open-weight language model generate text is not the same as making it finish coding work. This retrospective empirical systems experience report follows my engineering wo...
What Effort Level Should I Choose? On the Economics of Reasoning Effort in LLM Agents
Generative models increasingly expose a reasoning effort setting that lets users scale inference-time compute, trading higher cost for potentially improved performance. Users must ...
ChronoReplay: One Answer Interface Is Not Enough for Streaming Spatial Intelligence
Training a spatial VLM to answer which side an object lies on does not require it to preserve the metric displacement that determines that answer, while supervision for a distance ...
PreView: Enhancing GUI Grounding with Preprocessed Views and Cues
GUI grounding requires locating small controls in cluttered screenshots and identifying which control a natural-language instruction refers to. Cropping and enlargement can expose ...
Build, Don't Play: Self-Improving LLM Research Agents Develop a Hearthstone Battlegrounds AI That Rivals Top Players
Can LLM agents carry a hard real-world problem from raw data to a system that holds its own against people? We study this in Hearthstone Battlegrounds, an eight-player auto-battler...
Novel or Not? PatEx-Gym: Agentic Prior-Art Search for Novelty Assessment in Patent Examination
Patent examination is a challenging real-world task that requires agents to discover sparse prior art, iteratively refine queries, inspect technical evidence, and produce novelty a...
Trust Mechanisms for Offline Generative AI at the Edge: A Connectivity-Regime Framework and an Empirical Study on a Raspberry Pi 5
Generative AI and large language models (LLMs) are moving into the Internet of Things, yet most deployments assume cloud connectivity, and many IoT settings have intermittent, rest...
After One-Click Video: The Inflow of Service-Type Content, the Failure of the Recommendation Pool Mechanism, and the BoLe System
In September 2026, a new large language model was released. Soon after, a new way of making explainer videos became very popular: people asked the model to write a program, and the...
AV-FABLE: Audio-Visual Manipulation Forensics in Long-Form Videos
Localized AI manipulations in long-form videos often affect brief events within otherwise authentic content, making video-level classification insufficient for forensic analysis. E...
ReGE: Stronger Agents through Experience Synthesis
Repeated attempts at interactive tasks uncover facts, useful operations, and failures that can inform subsequent execution. We introduce ReGE, a framework for turning this recorded...
Semantic Concept Steering Breaks the Explanation Drift Loop in Continual Learning
Catastrophic forgetting in continual learning (CL) manifests not only as accuracy degradation but also as explanation drift: models maintain predictive accuracy while silently shif...