7354 papers
This paper introduces R2M, a novel lightweight RLHF framework that mitigates reward overoptimization by leveraging real-time policy hidden states to dynamically align the reward model with the policy's evolving distribution, rather than relying solely on static semantic representations.
This paper proposes a Bitcoin price prediction model using Combinatorial Fusion Analysis (CFA) to integrate diverse machine learning models via rank-score characteristics and weighted combinations, achieving a superior Mean Absolute Percentage Error (MAPE) of 0.19% that outperforms individual models and existing prediction methods.
This paper evaluates the impact of LLM-based news sentiment analysis on stock price prediction, demonstrating that DeBERTa outperforms other models and that an ensemble approach achieves 80% accuracy, while sentiment features provide modest improvements to various time-series forecasting architectures.
This paper introduces Adam-Aware In-Run Data Shapley, a novel method that overcomes the limitations of SGD-based attribution in adaptive optimizers by deriving a closed-form approximation and a Linearized Ghost Approximation to achieve near-perfect fidelity in data contribution estimation while maintaining high training efficiency.
This paper investigates whether Schwartz higher-order values improve sentence-level human value detection, finding that while hierarchical gating offers limited benefits, calibration techniques and hybrid ensembles significantly boost performance, suggesting the value hierarchy is more effective as an inductive bias than a rigid routing mechanism.
This paper introduces T2T (Thickening-to-Thinning), a dynamic reward shaping framework inspired by human learning dynamics that enhances LLM reasoning by encouraging longer, exploratory trajectories on incorrect attempts and penalizing length upon success, thereby outperforming standard baselines on mathematical benchmarks.
This paper introduces FSTab, a framework that demonstrates how LLM-generated software exhibits predictable, recurring vulnerabilities by enabling black-box attacks based on frontend features and quantifying the consistency of these flaws across different domains and model variations.
This paper introduces a scalable semantic search system for a corpus of 9.2 million mathematical theorems, demonstrating that representing theorems with natural-language descriptions significantly improves retrieval accuracy for both specific theorems and entire papers compared to existing baselines.
This paper introduces LMMRec, a model-agnostic framework that leverages large language models and chain-of-thought prompting to extract fine-grained user and item motivations from heterogeneous text data, effectively aligning them with interaction signals to significantly improve multimodal recommendation performance.
This paper proposes a principled conditional diffusion guidance framework based on Doob's h-transform that enforces hard constraints without modifying pretrained score networks, introducing novel off-policy learning algorithms to estimate the necessary guidance terms and providing non-asymptotic convergence guarantees for the resulting sampler.
This paper introduces NAAMSE, an evolutionary framework that automates and enhances AI agent security evaluation by using a single autonomous agent to iteratively mutate prompts and explore corpora, thereby uncovering adaptive vulnerabilities missed by static benchmarks while ensuring the models maintain benign-use correctness.
This paper proposes a unified diffusion-guided pretraining framework for brain graph foundation models that overcomes the limitations of existing methods by using diffusion to preserve semantic connectivity patterns during augmentation and to enable topology-aware global reconstruction, thereby achieving robust and transferable representations across diverse neuroimaging datasets.
The paper proposes CoCoA, a training-free decoding algorithm that mitigates LLM hallucinations by detecting and penalizing factually incorrect outputs through the analysis of representational instability and internal disagreement across the model's middle layers.
This paper proposes SToRM, a novel framework that employs a lightweight importance predictor, supervised training with pseudo-labels, and an anchor-context merging module to significantly reduce visual token redundancy in multi-modal LLMs for autonomous driving, achieving up to 30x computational savings while maintaining end-to-end performance comparable to using all tokens.
This paper introduces Agent-guided Policy Search (AGPS), a framework that replaces human supervisors with a multimodal agent acting as a semantic world model to provide precise corrective guidance, thereby significantly improving sample efficiency and scalability in robotic reinforcement learning compared to traditional Human-in-the-Loop methods.
This paper introduces M2RL, a comprehensive study comparing mixed multi-task training versus separate training with model merging for multi-domain Reinforcement Learning with Verifiable Rewards (RLVR), revealing that reasoning-intensive domains exhibit synergistic effects with minimal interference and providing mechanistic insights through extensive experiments.
The paper introduces SkillsBench, a comprehensive benchmark demonstrating that while curated agent skills significantly boost LLM performance across diverse domains—often allowing smaller models to match larger ones—self-generated skills offer no benefit and effects vary widely by task.
This paper proposes a geometric taxonomy of LLM hallucinations into three distinct types (unfaithfulness, confabulation, and factual error) and introduces corresponding detection metrics, the Semantic Grounding Index and Directional Grounding Index, which effectively identify unfaithful and confabulated outputs while revealing that apparent signals for factual errors in existing benchmarks often stem from stylistic annotation confounds rather than genuine geometric distinctions.
TrasMuon is a novel optimizer that enhances the stability and convergence of Muon-style methods by integrating global RMS calibration and energy-based trust-region clipping to preserve near-isometric geometry while mitigating sensitivity to step-size hyperparameters and high-energy outliers.
This paper demonstrates that a lightweight, automated AI pipeline integrating next-generation large language models with citation-based verification can successfully generate and solve sophisticated, research-grade mathematical problems, including previously unpublished questions, with verified results and open-sourced tools.