Leaderboards, eval harnesses and ablations — the contested business of deciding which model is actually better.
arXiv:2609.38536v1 Announce Type: cross Abstract: Diffusion-based large language models (dLLMs) promise to break the sequential latency bottleneck of autoregressive agents through parallel decoding,...
arXiv:2609.38547v1 Announce Type: cross Abstract: Defining a weighted mean over probability measures under probability metrics is a central tool in probabilistic machine learning. Under the Wasserste...
ChartDensity-Bench is a benchmark designed to evaluate multimodal large language models (MLLMs) on their ability to reconstruct structured numerical data from scientific charts that vary in visual density. The benchmark uses charts paired with source-level ground-truth data and systematically changes the number of simultaneously presented charts (k = 1, 3, 6, 9) to assess how density affects reconstruction performance. A multi‑dimensional evaluation framework measures structural reliability, reconstruction completeness, parseability, and numerical fidelity, revealing that numerical reconstruction generally worsens as visual density increases, with varying degrees of degradation across different models.
arXiv:2609.38863v1 Announce Type: cross Abstract: The two-dimensional irregular knapsack problem in a fixed circular container is an important combinatorial optimization problem for maximizing materi...
arXiv:2609.38879v1 Announce Type: cross Abstract: Large language models rely heavily on human text, which often conveys surface answers rather than the spatial and structural logic behind them. Prote...
The paper introduces OptTips, a knowledge base of 50 expert optimization modeling techniques, and OptDachshund, a multi‑agent framework that generates mathematical models and solver code from natural‑language problem descriptions. Using these tools, the authors create the EfficientOpt benchmark, comprising 561 expert‑reviewed tasks with paired reference implementations, to evaluate large language models (LLMs) on both correctness and computational efficiency. Their experiments with 11 LLMs show a consistent efficiency gap: even when LLMs produce correct solutions, the resulting programs often take longer to solve than expert‑crafted counterparts, highlighting the need to assess both accuracy and runtime performance in LLM‑based optimization modeling.
The paper introduces PACT, a method for unlearning deceptive behaviors in large language models by using contrastive forget sets that compare a model’s responses under deceptive and neutral contexts. PACT trains the model to produce pressure‑aware counterfactual targets, preserving benign system‑prompt adherence and reasoning traces while dramatically reducing deception rates from over 50% to under 3% on 32B reasoning models.
arXiv:2609. 38977v1 Announce Type: cross Abstract: Neural surrogates have emerged as fast alternatives to the numerical simulation of three-dimensional turbulence.
arXiv:2609.39257v1 Announce Type: cross Abstract: Accurate forecasting of electricity production is essential for maintaining the operational efficiency and strategic planning of energy utilities. In...
arXiv:2609.39321v1 Announce Type: cross Abstract: Group Relative Policy Optimization (GRPO) has emerged as a memory-efficient reinforcement fine-tuning (RFT) technique for reasoning-intensive tasks....
The paper introduces GFD-OPD, a method for on‑policy distillation of diffusion models that addresses challenges when compressing large teachers into smaller students. It identifies that standard distillation fails due to distribution gaps and classifier‑free guidance amplification, and proposes Fixed‑State KL to measure these gaps. GFD‑OPD reduces the student‑teacher discrepancy and achieves state‑of‑the‑art performance across multiple benchmarks.
The paper introduces probabilistic adversarial training, a method that enhances robustness by reducing the overlap between a distance-based distribution and a victim-classifier-induced distribution. It derives a KL-based lower bound on probabilistic robustness, which serves as a tractable surrogate objective. Experiments confirm that this approach consistently improves probabilistic robustness and can even boost non-probabilistic adversarial training methods through an induced scaling factor.
arXiv:2609.39934v1 Announce Type: cross Abstract: Checkpoint selection in domain generalization often relies on source-validation accuracy, yet the selected checkpoint need not provide reliable proba...
The paper proposes an inference auction for large language model (LLM) APIs, enabling users to bid for faster service when compute demand exceeds capacity. The auction allocates priority efficiently without increasing latency, and includes fast algorithms for truthful bidding and an autobidding agent that adjusts bids within a user’s budget to maximize utility. Experiments show the auction improves system welfare while preserving the cache utilization and latency benefits of the SGLang inference framework.
arXiv:2609.40316v1 Announce Type: cross Abstract: Looped transformers and Mixture-of-Experts (MoE) offer complementary routes to efficient scaling: recurrence increases computational depth at fixed p...
arXiv:2609.38274v1 Announce Type: new Abstract: The performance ceiling of an LLM team is constrained not only by individual model capabilities, but also by inter-member error resonance and predictiv...
The paper introduces a self‑evolving harness framework where a frozen language‑model agent first solves tasks and then edits its own harness based on run records. Using a 49‑line seed harness, the evolved harness improves average scores on in‑distribution benchmarks by 4.48 points and on out‑of‑distribution benchmarks by 12.64 points, surpassing Codex on the former and matching it on the latter. Continued evolution on a specific out‑of‑distribution benchmark further raises performance, and the study analyzes emergent mechanisms such as output truncation and history compaction.
GazeFlow is a new framework for egocentric gaze prediction that models gaze as a joint distribution of temporal positions conditioned on both top‑down task cues and bottom‑up visual saliency. It employs conditional flow matching to iteratively transform Gaussian noise into realistic gaze trajectories, using a velocity field informed by video‑encoded visual features and global task queries. On standard benchmarks, GazeFlow outperforms existing methods on per‑frame metrics and produces trajectories that better reflect human gaze dynamics.
arXiv:2609.38530v1 Announce Type: new Abstract: Language models increasingly use architectures that vary attention span and positional encoding across layers, such as applying RoPE with sliding-windo...
arXiv:2609.38543v1 Announce Type: new Abstract: Constantly evolving real-world knowledge necessitates models to be updated continuously. Especially in medicine, as clinical evidence changes over time...