The advent of large language models (LLMs), such as GPT, Gemini, and DeepSeek, has significantly advanced natural language processing, giving rise to sophisticated chatbots capable of diverse language-related tasks. The transition from these traditional LLM chatbots to more advanced AI agents represents a pivotal evolutionary step. However, existing evaluation frameworks often blur the distinctions between LLM chatbots and AI agents, leading to confusion among researchers selecting appropriate benchmarks. To bridge this gap, this paper introduces a systematic analysis of current evaluation approaches, grounded in an evolutionary perspective. We provide a detailed analytical framework that clearly differentiates AI agents from LLM chatbots along five key aspects: complex environment, multi-source instructor, dynamic feedback, multi-modal perception, and advanced capability. Further, we categorize existing evaluation benchmarks based on external environments driving forces, and resulting advanced internal capabilities. For each category, we delineate relevant evaluation attributes, presented comprehensively in practical reference tables. Finally, we synthesize current trends and outline future evaluation methodologies through four critical lenses: environment, agent, evaluator, and metrics. Our findings offer actionable guidance for researchers, facilitating the informed selection and application of benchmarks in AI agent evaluation, thus fostering continued advancement in this rapidly evolving research domain.
Charge-parity (CP) violation in the tau-charm energy region is a promising area for sensitive tests of Standard Model (SM) predictions and searches for new, beyond the SM physics. A future Tau-Charm Facility that operates at center-of-mass energies between 2.0 and 7.0 GeV, with a peak luminosity of 0.5×1035 cm−2 s−1, would provide huge numbers of hadrons and tau (τ) leptons that are produced in low-background environments and with well understood kinematic properties. In this report, prospects for unique studies of CP violation in the decay of charmed hadrons, and in the production and decay of hyperons and τ leptons at a next-generation tau-charm facility are discussed. In addition, opportunities for improved tests of CPT invariance test in K0−K̄0 mixing are presented.
Existing legal benchmarks for Large Language Models (LLMs) primarily focus on final judgment outcomes or isolated subtasks, making it difficult to systematically evaluate the multi-step reasoning processes and reasoning quality required in real-world legal decision-making. To address this gap, we introduce MSLR, a Chinese benchmark for multi-step legal reasoning, which models and evaluates complete judicial reasoning trajectories based on the IRAC (Issue-Rule-Application-Conclusion) framework. MSLR comprises 1389 Chinese insider trading judgments issued between 2005 and 2024, with approximately 60,000 step-level reasoning paths annotated through a combination of human and automated methods, averaging 43 intermediate reasoning steps per case. Based on MSLR, we adopt two evaluation metrics, IRAC Recall and LLM Score, to provide a fine-grained assessment of legal reasoning capabilities. We systematically evaluate 19 mainstream general-purpose, legal-domain, and reasoning-oriented models. The results show that current models exhibit overall limited performance on multi-step legal reasoning tasks; even the best-performing o1-mini model fails to exceed 75% IRAC Recall. In addition, we report the performance of mainstream models on structured annotation tasks and legal judgment reasoning tasks. The MSLR dataset and benchmark construction code are publicly available at https://github.com/yuwenhan07/MSLR-Bench.
Modern serving systems for Mixture-of-Experts (MoE) models adopt hybrid data-expert parallelism: expert parallelism (EP) shards experts across GPUs to scale capacity, while data parallelism (DP) replicates attention layers across instances to process independent requests. Existing systems bind each request's attention, MoE communication, and KV cache to a single instance. Because attention latency scales with KV cache size while MoE communication latency scales with batch size, this binding cannot balance both simultaneously, producing EP stragglers; it also fragments KV memory across instances, inflating tail latency under long contexts. While existing context parallelism (CP) mitigates these constraints, existing group-level and uniform CP designs incur prohibitive communication and attention-side overheads. We present NanoCP, which decouples MoE communication from KV cache placement and achieves dual balance through dynamic context parallelism (DCP). DCP assigns each request a context-parallel degree sized to its KV footprint: long requests distribute attention across multiple instances; short requests remain local. This dynamic parallelism effectively liquefies the KV cache across DP instances within each node, balancing both the per-instance KV cache occupancy and batch sizes without unnecessary load-balancing costs. To bridge DCP with static execution, NanoCP introduces an ahead-of-time (AOT) graph engine paired with a custom routing-based communication backend. Experimental results show that, under 99
The high-voltage spinel LiNi0.5 Mn1.5 O4 (LNMO) is a promising cobalt-free cathode with high specific energy, attributed to its high operating voltage (approximately 4.7 V). However, its practical durability is limited by irreversible transition-metal migration, spinel-to-rock-salt phase reconstruction, highvoltage interfacial side reactions, and transition-metal dissolution. Herein, we report a multi-elementdoped, cation-disordered spinel LiNi0.4 Mn1.4 Fe0.05 Cr0.05 Cu0.05 Nb0.05 O4 (HELNMO) synthesized via a solidstate method. Rietveld refinement confirms a single-phase Fd 3 m structure with a slightly expanded lattice parameter (8.1857 A & ring; for HELNMO versus 8.1776 A & ring; for LNMO), and energy-dispersive X-ray spectroscopy mapping reveals a homogeneous distribution of the dopant elements. Structural and spectroscopic characterizations consistently confirm enhanced cation disorder in HELNMO, which contributes to favorable electrochemical performance by delivering approximately 130 mAh g-1 at 0.1 C and exhibiting markedly improved cycling stability, retaining 92 % of its initial capacity after 10 0 0 cycles at 2 C between 3.5 and 4.9 V. In-situ X-ray diffraction reveals smoother, more continuous lattice evolution of HELNMO than LNMO during (de)lithiation, consistent with a more homogeneous reaction pathway. Post-cycling analyses corroborate suppressed surface rock-salt reconstruction, reduced impedance growth, and alleviated Mn/Ni dissolution. Collectively, multi-element doping coupled with disorder engineering can effectively mitigate transition-metal migration and high-voltage interfacial degradation, providing a practical strategy for durable, high-energy spinel cathodes. (c) 2026 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.
