ESPO: Early-Stopping Proximal Policy Optimization
When a large language model under reinforcement learning commits a wrong reasoning step early in a trajectory, standard algorithms force it to keep generating until the maximum hor...
Papers
Reducing Political Manipulation with Consistency Training
Large language models (LLMs) exhibit systematic political bias across a variety of sensitive contexts. We find that LLMs handle counterpart topics from opposing political sides asy...
Access Sets Matter: Budgeting Expert Reads for Scalable Weight-Space Model Merging
Weight-space model merging is usually formulated as an algebraic operation on checkpoints, yet at LLM scale the limiting resource is often the set of expert weights that must be re...
Seeing Isn't Knowing: Do VLMs Know When Not to Answer Spatial Questions (and Why)?
Spatial reasoning is a fundamental capability for vision-language models (VLMs) deployed in real-world environments. However, visual observations are inherently limited representat...
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
LLM agents are increasingly deployed as systems built around editable external harnesses, including prompts, skills, memories and tools, that shape task execution without changing ...
Prior Availability in Industrial Visual Sim-to-Real: A Review of CAD-Guided and CAD-Unavailable Regimes
Industrial visual sim-to-real is often described as transferring from synthetic images to real images, but industrial deployment usually involves a broader mismatch between availab...
Linear Ensembles Wash Away Watermarks: On the Fragility of Distributional Perturbations in LLMs
Watermarking embeds statistical signatures in AI-generated text for detection and attribution. We reveal a fundamental vulnerability: when users access multiple models (today's rea...
SoundnessBench: Can Your AI Scientist Really Tell Good Research Ideas from Bad Ones?
Autonomous AI research agents aim to accelerate scientific discovery by automating the research pipeline, from hypothesis generation to peer review. However, existing benchmarks ra...
Pressure-Testing Deception Probes in LLMs: Scaling, Robustness, and the Geometry of Deceptive Representations
Linear probes trained on LLM activations are increasingly proposed as deception-detection metrics, yet report AUROC exceeding 0.96 on clean benchmarks while collapsing under distri...
OR-Space: A Full-Lifecycle Workspace Benchmark for Industrial Optimization Agents
Large language model (LLM) agents are increasingly used to assist with operations research (OR) modeling, yet existing OR-oriented benchmarks often reduce evaluation to one-shot tr...
Efficient and Scalable Provenance Tracking for LLM-Generated Code Snippets
Large language models (LLMs) for code completion and generation are increasingly used in software development, yet they may reproduce training examples verbatim and without authors...
RUBRIC-ARROW: Alternating Pointwise Rubric Reward Modeling for LLM Post-training in Non-verifiable Domains
Pointwise reward modeling offers critical signals for LLM post-training, yet struggles with absolute scoring in subjective, non-verifiable settings. Rubric-based methods address th...
PEFT-Arena: Understanding Parameter-Efficient Finetuning from a Stability-Plasticity Perspective
Parameter-efficient finetuning (PEFT) has become the standard approach for adapting large language models, yet evaluations largely emphasize downstream accuracy while overlooking t...
ProRL: Effective Reinforcement Learning for Proactive Recommendation via Rectified Policy Gradient Estimation
Proactive Recommender Systems (PRSs) aim to guide user preference shift toward target items by generating paths of intermediate recommendations. Reinforcement learning (RL) provide...
Beyond Recall: Behavioral Specification as an Interpretive Layer for AI Personalization
If an AI agent makes decisions on a person's behalf, those decisions must align with its user. We introduce representational accuracy to measure how faithfully a system captures a ...
DEMON: Diffusion Engine for Musical Orchestrated Noise
We present DEMON, a real-time diffusion engine that makes the denoising process playable as a live musical instrument: a control surface both broad (many parameters shaped per-fram...
The Hamilton-Jacobi Theory of Deep Learning
In this paper, training a neural network is identified, exactly, as a search through Hamilton--Jacobi initial-value problems: each gradient step selects the initial data of a visco...
CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning
Language models can use verifiable rewards to improve at a wide variety of reasoning tasks. However, both parametric (e.g. RLVR) and non-parametric (e.g. prompt optimization) appro...
PRISM: A Multi-Dimensional Benchmark for Evaluating LLM Peer Reviewers
The rapid growth in submissions to machine learning venues has strained the scientific peer-review system and intensified interest in LLM-based automated peer reviewers. However, h...
AI Research Agents Narrow Scientific Exploration
AI research agents can now generate research ideas, design experiments, run code, and draft papers, raising the possibility of large-scale AI-assisted scientific discovery. Many cu...
Joint Training of Multi-Token Prediction in Reinforcement Learning via Optimal Coefficient Calibration
Reinforcement Learning from Verifiable Rewards (RLVR) has emerged as the standard paradigm for improving reasoning capability of large language models, while Multi-Token Prediction...
Revealing Algorithmic Deductive Circuits for Logical Reasoning
Recent studies have shown that Large Language Models (LLMs) can achieve strong reasoning performance by incorporating functional symbolic representations that abstractly describe g...
Learn from Weaknesses: Automated Domain Specialization for Small Computer-Use Agents
Computer-use agents (CUAs) have recently made substantial progress, but deploying a separate large expert for each software domain remains expensive. Small open computer-use agents...
Clark Hash: Stateless Sparse Johnson-Lindenstrauss Quantization for Neural Embeddings
Clark Hash is a small method for storing neural embeddings in less space. It normalizes each database vector, applies a deterministic sparse signed Johnson-Lindenstrauss projection...