Leaderboards, eval harnesses and ablations — the contested business of deciding which model is actually better.
arXiv:2605.12922v2 Announce Type: replace Abstract: Large language models can follow complex instructions in a single turn, yet over long multi-turn interactions they often lose the thread of instruc...
arXiv:2605.13335v2 Announce Type: replace Abstract: Embodied agents in household environments must plan under partial observation: they need to remember objects, track state changes, and recover when...
arXiv:2605.24154v2 Announce Type: replace Abstract: Current safety alignment of foundation models largely follows a \emph{one-size-fits-all} paradigm, applying the same refusal policy across users an...
arXiv:2606.21399v2 Announce Type: replace Abstract: Runtime oversight often intervenes when an LLM agent's calibrated failure score crosses a threshold. Yet states with the same failure risk can diff...
SWE-Game is a benchmark comprising 247 tasks based on 41 Godot games across 13 gameplay categories, testing coding agents on tasks such as brief-to-game, design-document implementation, skeleton completion, fault repair, and Godot-to-Unity porting. Evaluation uses engine-state checks, replay of certified reference inputs, and agent-authored demonstrations to judge mechanic correctness, playability, and post-repair behavior, supplemented by vision‑language rubrics for presentation. Across six models, Opus5 leads but overall scores stay below 60/100, highlighting common issues like omitted requirements and gameplay logic errors.
arXiv:2610.03387v2 Announce Type: replace Abstract: Rejection policies must remain useful as candidate sets and tasks change. We compare Laya, Jev and Qwen2.5-7B-Instruct using public reference label...
arXiv:2610.05370v2 Announce Type: replace Abstract: Generative reward models (GRMs) are important for LLM optimization. Unlike scalar reward models, GRMs generate natural-language critiques alongside...
arXiv:2610.05437v2 Announce Type: replace Abstract: Computer-use agents need to capture procedural knowledge of how people use software. User telemetry offers a scalable source of this knowledge. How...
arXiv:2610.06824v2 Announce Type: replace Abstract: We introduce TasteVal, a benchmark to evaluate the experimental research taste of frontier models. We define research taste as the ability to pick...
arXiv:2311.18029v2 Announce Type: replace-cross Abstract: The current trend in the literature on Time Series Classification is to develop increasingly accurate algorithms by combining multiple models...
arXiv:2402.02399v3 Announce Type: replace-cross Abstract: Time series modeling presents unique challenges due to autocorrelation in both historical data and future sequences. While current research p...
arXiv:2408.12549v4 Announce Type: replace-cross Abstract: This paper presents a method for modeling optical dynamic range compressors using deep neural networks with Selective State Space models. The...
arXiv:2503.22764v3 Announce Type: replace-cross Abstract: The large language model (LLM) is typically integrated into the mainstream optimization protocol. However, it remains underexplored whether m...
arXiv:2505.14777v2 Announce Type: replace-cross Abstract: The design of effective optimization algorithms for neural networks remains a fundamental challenge, and most existing methods rely on heuris...
arXiv:2508.05880v3 Announce Type: replace-cross Abstract: Understanding human emotions is central to user-facing AI applications, safety alignment, and the simulation of human behavior. As emotional...
arXiv:2508.10020v2 Announce Type: replace-cross Abstract: Enhancing LLM reasoning in federated settings is nontrivial due to stringent computational, communication, and privacy constraints, especiall...
arXiv:2510.11593v3 Announce Type: replace-cross Abstract: For reliable large-scale quantum computation, quantum error correction (QEC) is essential to protect logical information distributed across m...
arXiv:2511.00053v2 Announce Type: replace-cross Abstract: The design of learning objectives is central to training time-series forecasting models. Existing learning objectives such as mean squared er...
The paper introduces SORL, a framework that stabilizes off‑policy reinforcement learning for long‑horizon large language model agents. It identifies two key instability sources—token‑level policy granularity mismatches and high‑variance off‑policy updates—and proposes turn‑level importance sampling and clipping‑triggered normalization to align optimization with multi‑turn interactions. Two instantiations, SO‑PPO and SO‑GRPO, are evaluated on open‑domain, multi‑hop, and medical QA benchmarks, as well as on asynchronous RL for mathematical reasoning, showing improved robustness without the need for early stopping or heuristic tuning.
arXiv:2603.01295v2 Announce Type: replace-cross Abstract: Joint lesion segmentation and tissue classification in breast ultrasound are usually trained with a shared encoder, so the two branches stop...