Daily Papers

We adapt Microsoft's QuantumKatas -- a well-established quantum computing curriculum -- from Q# to Qiskit, the most widely-adopted quantum computing framework, and package it with an evaluation framework for systematic LLM assessment. The resulting benchmark comprises 350 tasks across 26 categories, spanning fundamental gates through advanced algorithms (Grover's, Simon's, Deutsch-Jozsa), error correction, key distribution, and quantum games. Each task includes a natural language prompt, canonical solution, and deterministic test verification via classical circuit simulation. By building on the QuantumKatas' proven pedagogical design rather than creating tasks from scratch, we inherit a principled difficulty progression and comprehensive concept coverage while contributing the framework adaptation, evaluation infrastructure, and empirical analysis. We evaluate 16 LLMs across 7 prompting configurations -- a total of 39,200 model runs -- to demonstrate the benchmark's utility. Three key findings emerge: (1) the benchmark effectively differentiates model capabilities, with best-configuration pass rates ranging from 32.3% to 83.1% and a 26.1 pp average gap between frontier and open-source models; (2) models perform well at implementing known algorithms (SimonsAlgorithm 82.1%, BasicGates 81.6%) but struggle with problem encoding (SolveSATWithGrover 34.4%, DistinguishUnitaries 40.0%); and (3) chain-of-thought prompting shows a modestly bimodal effect -- it is the best strategy for three models (two of them explicitly reasoning-tuned per vendor documentation) but degrades performance for the rest, leaving it mid-pack in aggregate (56.3% mean) behind few-shot-5 (57.8%). We release the benchmark, evaluation framework, and baseline results to support research on LLM capabilities in quantum computing.

Large Language Models (LLMs) are becoming increasingly smart and autonomous, targeting real-world pragmatic missions beyond traditional NLP tasks. As a result, there has been an urgent need to evaluate LLMs as agents on challenging tasks in interactive environments. We present AgentBench, a multi-dimensional evolving benchmark that currently consists of 8 distinct environments to assess LLM-as-Agent's reasoning and decision-making abilities in a multi-turn open-ended generation setting. Our extensive test over 25 LLMs (including APIs and open-sourced models) shows that, while top commercial LLMs present a strong ability of acting as agents in complex environments, there is a significant disparity in performance between them and open-sourced competitors. It also serves as a component of an ongoing project with wider coverage and deeper consideration towards systematic LLM evaluation. Datasets, environments, and an integrated evaluation package for AgentBench are released at https://github.com/THUDM/AgentBench

Evaluating Large Language Models (LLMs) for mental health support is challenging due to the emotionally and cognitively complex nature of therapeutic dialogue. Existing benchmarks are limited in scale, reliability, often relying on synthetic or social media data, and lack frameworks to assess when automated judges can be trusted. To address the need for large-scale dialogue datasets and judge reliability assessment, we introduce two benchmarks that provide a framework for generation and evaluation. MentalBench-100k consolidates 10,000 one-turn conversations from three real scenarios datasets, each paired with nine LLM-generated responses, yielding 100,000 response pairs. MentalAlign-70k}reframes evaluation by comparing four high-performing LLM judges with human experts across 70,000 ratings on seven attributes, grouped into Cognitive Support Score (CSS) and Affective Resonance Score (ARS). We then employ the Affective Cognitive Agreement Framework, a statistical methodology using intraclass correlation coefficients (ICC) with confidence intervals to quantify agreement, consistency, and bias between LLM judges and human experts. Our analysis reveals systematic inflation by LLM judges, strong reliability for cognitive attributes such as guidance and informativeness, reduced precision for empathy, and some unreliability in safety and relevance. Our contributions establish new methodological and empirical foundations for reliable, large-scale evaluation of LLMs in mental health. We release the benchmarks and codes at: https://github.com/abeerbadawi/MentalBench/

The increasing context length of modern language models has created a need for evaluating their ability to retrieve and process information across extensive documents. While existing benchmarks test long-context capabilities, they often lack a structured way to systematically vary question complexity. We introduce KG-QAGen (Knowledge-Graph-based Question-Answer Generation), a framework that (1) extracts QA pairs at multiple complexity levels (2) by leveraging structured representations of financial agreements (3) along three key dimensions -- multi-hop retrieval, set operations, and answer plurality -- enabling fine-grained assessment of model performance across controlled difficulty levels. Using this framework, we construct a dataset of 20,139 QA pairs (the largest number among the long-context benchmarks) and open-source a part of it. We evaluate 13 proprietary and open-source LLMs and observe that even the best-performing models are struggling with set-based comparisons and multi-hop logical inference. Our analysis reveals systematic failure modes tied to semantic misinterpretation and inability to handle implicit relations.

As the demand for comprehensive evaluations of diverse model capabilities steadily increases, benchmark suites have correspondingly grown significantly in scale. Despite notable advances in redundancy reduction and subset-level performance prediction, a systematic framework that effectively integrates these methods to ensure both prediction accuracy and ranking consistency is still largely elusive. In this paper, we first perform a sample-level analysis of benchmark redundancy and identify several highly similar samples that can be eliminated. Besides, we frame benchmark compression as an optimization problem with the aim of score reconstruction. Building on these, we then propose EssenceBench, a coarse-to-fine framework utilizing an iterative Genetic Algorithm (GA), which takes the advantages of fitness-based subset search and attribution-based sample search. Compared to previous methods, our approach yields superior compression results with lower reconstruction error and markedly higher efficiency. In particular, on the HellaSwag benchmark (10K samples), our method preserves the ranking of all models shifting within 5% using 25x fewer samples, and achieves 95% ranking preservation shifting within 5% using only 200x fewer samples.

Large language models (LLMs) are increasingly evaluated and sometimes trained using automated graders such as LLM-as-judges that output scalar scores or preferences. While convenient, these approaches are often opaque: a single score rarely explains why an answer is good or bad, which requirements were missed, or how a system should be improved. This lack of interpretability limits their usefulness for model development, dataset curation, and high-stakes deployment. Query-specific rubric-based evaluation offers a more transparent alternative by decomposing quality into explicit, checkable criteria. However, manually designing high-quality, query-specific rubrics is labor-intensive and cognitively demanding and not feasible for deployment. While previous approaches have focused on generating intermediate rubrics for automated downstream evaluation, it is unclear if these rubrics are both interpretable and effective for human users. In this work, we investigate whether LLMs can generate useful, instance-specific rubrics as compared to human-authored rubrics, while also improving effectiveness for identifying good responses. Through our systematic study on two rubric benchmarks, and on multiple few-shot and post-training strategies, we find that off-the-shelf LLMs produce rubrics that are poorly aligned with human-authored ones. We introduce a simple strategy, RubricRAG, which retrieves domain knowledge via rubrics at inference time from related queries. We demonstrate that RubricRAG can generate more interpretable rubrics both for similarity to human-authored rubrics, and for improved downstream evaluation effectiveness. Our results highlight both the challenges and a promising approach of scalable, interpretable evaluation through automated rubric generation.

Personality manipulation in large language models (LLMs) is increasingly applied in customer service and agentic scenarios, yet its mechanisms and trade-offs remain unclear. We present a systematic study of personality control using the Big Five traits, comparing in-context learning (ICL), parameter-efficient fine-tuning (PEFT), and mechanistic steering (MS). Our contributions are fourfold. First, we construct a contrastive dataset with balanced high/low trait responses, enabling effective steering vector computation and fair cross-method evaluation. Second, we introduce a unified evaluation framework based on within-run Delta analysis that disentangles, reasoning capability, agent performance, and demographic bias across MMLU, GAIA, and BBQ benchmarks. Third, we develop trait purification techniques to separate openness from conscientiousness, addressing representational overlap in trait encoding. Fourth, we propose a three-level stability framework that quantifies method-, trait-, and combination-level robustness, offering practical guidance under deployment constraints. Experiments on Gemma-2-2B-IT and LLaMA-3-8B-Instruct reveal clear trade-offs: ICL achieves strong alignment with minimal capability loss, PEFT delivers the highest alignment at the cost of degraded task performance, and MS provides lightweight runtime control with competitive effectiveness. Trait-level analysis shows openness as uniquely challenging, agreeableness as most resistant to ICL, and personality encoding consolidating around intermediate layers. Taken together, these results establish personality manipulation as a multi-level probe into behavioral representation, linking surface conditioning, parameter encoding, and activation-level steering, and positioning mechanistic steering as a lightweight alternative to fine-tuning for both deployment and interpretability.

We introduce Debate Speech Evaluation as a novel and challenging benchmark for assessing LLM judges. Evaluating debate speeches requires a deep understanding of the speech at multiple levels, including argument strength and relevance, the coherence and organization of the speech, the appropriateness of its style and tone, and so on. This task involves a unique set of cognitive abilities that have previously received limited attention in systematic LLM benchmarking. To explore such skills, we leverage a dataset of over 600 meticulously annotated debate speeches and present the first in-depth analysis of how state-of-the-art LLMs compare to human judges on this task. Our findings reveal a nuanced picture: while larger models can approximate individual human judgments in some respects, they differ substantially in their overall judgment behavior. We also investigate the ability of frontier LLMs to generate persuasive, opinionated speeches, showing that models may perform at a human level on this task.

LLM-as-Judge frameworks are increasingly popular for AI evaluation, yet research findings on the relationship between models' generation and judgment abilities remain inconsistent. We investigate this relationship through systematic dataset- and instance-level analyses across 11 models and 21 diverse tasks. Despite both capabilities relying on the same underlying knowledge, our analyses reveal they are only weakly correlated, primarily due to LLMs' sensitivity to the responses being judged. To address this, we propose a self-reference-guided evaluation strategy that leverages a model's own answers as references. This approach significantly strengthens the correlation between generation and judgment abilities, offering a practical path to align these skills and providing a reliable proxy for model selection in evaluation tasks.

One of the most widely used tasks to evaluate Large Language Models (LLMs) is Multiple-Choice Question Answering (MCQA). While open-ended question answering tasks are more challenging to evaluate, MCQA tasks are, in principle, easier to assess, as the model's answer is thought to be simple to extract and is directly compared to a set of predefined choices. However, recent studies have started to question the reliability of MCQA evaluation, showing that multiple factors can significantly impact the reported performance of LLMs, especially when the model generates free-form text before selecting one of the answer choices. In this work, we shed light on the inconsistencies of MCQA evaluation strategies, which can lead to inaccurate and misleading model comparisons. We systematically analyze whether existing answer extraction methods are aligned with human judgment, and how they are influenced by answer constraints in the prompt across different domains. Our experiments demonstrate that traditional evaluation strategies often underestimate LLM capabilities, while LLM-based answer extractors are prone to systematic errors. Moreover, we reveal a fundamental trade-off between including format constraints in the prompt to simplify answer extraction and allowing models to generate free-form text to improve reasoning. Our findings call for standardized evaluation methodologies and highlight the need for more reliable and consistent MCQA evaluation practices.

We present QIMMA, a quality-assured Arabic LLM leaderboard that places systematic benchmark validation at its core. Rather than aggregating existing resources as-is, QIMMA applies a multi-model assessment pipeline combining automated LLM judgment with human review to surface and resolve systematic quality issues in well-established Arabic benchmarks before evaluation. The result is a curated, multi-domain, multi-task evaluation suite of over 52k samples, grounded predominantly in native Arabic content; code evaluation tasks are the sole exception, as they are inherently language-agnostic. Transparent implementation via LightEval, EvalPlus and public release of per-sample inference outputs make QIMMA a reproducible and community-extensible foundation for Arabic NLP evaluation.

Recent advancements in large language models (LLMs) have underscored their importance in the evolution of artificial intelligence. However, despite extensive pretraining on multilingual datasets, available open-sourced LLMs exhibit limited effectiveness in processing Vietnamese. The challenge is exacerbated by the absence of systematic benchmark datasets and metrics tailored for Vietnamese LLM evaluation. To mitigate these issues, we have finetuned LLMs specifically for Vietnamese and developed a comprehensive evaluation framework encompassing 10 common tasks and 31 metrics. Our evaluation results reveal that the fine-tuned LLMs exhibit enhanced comprehension and generative capabilities in Vietnamese. Moreover, our analysis indicates that models with more parameters can introduce more biases and uncalibrated outputs and the key factor influencing LLM performance is the quality of the training or fine-tuning datasets. These insights underscore the significance of meticulous fine-tuning with high-quality datasets in enhancing LLM performance.

Large Language Models (LLMs) have recently gained significant attention due to their remarkable capabilities in performing diverse tasks across various domains. However, a thorough evaluation of these models is crucial before deploying them in real-world applications to ensure they produce reliable performance. Despite the well-established importance of evaluating LLMs in the community, the complexity of the evaluation process has led to varied evaluation setups, causing inconsistencies in findings and interpretations. To address this, we systematically review the primary challenges and limitations causing these inconsistencies and unreliable evaluations in various steps of LLM evaluation. Based on our critical review, we present our perspectives and recommendations to ensure LLM evaluations are reproducible, reliable, and robust.

Recent advancements in large language models (LLMs) have given rise to the LLM-as-a-judge paradigm, showcasing their potential to deliver human-like judgments. However, in the field of machine translation (MT) evaluation, current LLM-as-a-judge methods fall short of learned automatic metrics. In this paper, we propose Multidimensional Multi-Agent Debate (M-MAD), a systematic LLM-based multi-agent framework for advanced LLM-as-a-judge MT evaluation. Our findings demonstrate that M-MAD achieves significant advancements by (1) decoupling heuristic MQM criteria into distinct evaluation dimensions for fine-grained assessments; (2) employing multi-agent debates to harness the collaborative reasoning capabilities of LLMs; (3) synthesizing dimension-specific results into a final evaluation judgment to ensure robust and reliable outcomes. Comprehensive experiments show that M-MAD not only outperforms all existing LLM-as-a-judge methods but also competes with state-of-the-art reference-based automatic metrics, even when powered by a suboptimal model like GPT-4o mini. Detailed ablations and analysis highlight the superiority of our framework design, offering a fresh perspective for LLM-as-a-judge paradigm. Our code and data are publicly available at https://github.com/SU-JIAYUAN/M-MAD.

Large Language Models for code (code LLMs) have witnessed tremendous progress in recent years. With the rapid development of code LLMs, many popular evaluation benchmarks, such as HumanEval, DS-1000, and MBPP, have emerged to measure the performance of code LLMs with a particular focus on code generation tasks. However, they are insufficient to cover the full range of expected capabilities of code LLMs, which span beyond code generation to answering diverse coding-related questions. To fill this gap, we propose InfiBench, the first large-scale freeform question-answering (QA) benchmark for code to our knowledge, comprising 234 carefully selected high-quality Stack Overflow questions that span across 15 programming languages. InfiBench uses four types of model-free automatic metrics to evaluate response correctness where domain experts carefully concretize the criterion for each question. We conduct a systematic evaluation for over 100 latest code LLMs on InfiBench, leading to a series of novel and insightful findings. Our detailed analyses showcase potential directions for further advancement of code LLMs. InfiBench is fully open source and continuously expanding to foster more scientific and systematic practices for code LLM evaluation.

Traditional AI safety evaluations on isolated LLMs are insufficient as multi-agent AI ensembles become prevalent, introducing novel emergent risks. This paper introduces the Multi-Agent Emergent Behavior Evaluation (MAEBE) framework to systematically assess such risks. Using MAEBE with the Greatest Good Benchmark (and a novel double-inversion question technique), we demonstrate that: (1) LLM moral preferences, particularly for Instrumental Harm, are surprisingly brittle and shift significantly with question framing, both in single agents and ensembles. (2) The moral reasoning of LLM ensembles is not directly predictable from isolated agent behavior due to emergent group dynamics. (3) Specifically, ensembles exhibit phenomena like peer pressure influencing convergence, even when guided by a supervisor, highlighting distinct safety and alignment challenges. Our findings underscore the necessity of evaluating AI systems in their interactive, multi-agent contexts.

LLM-as-a-Judge has been widely applied to evaluate and compare different LLM alignmnet approaches (e.g., RLHF and DPO). However, concerns regarding its reliability have emerged, due to LLM judges' biases and inconsistent decision-making. Previous research has developed evaluation frameworks to assess reliability of LLM judges and their alignment with human preferences. However, the employed evaluation metrics often lack adequate explainability and fail to address LLM internal inconsistency. Additionally, existing studies inadequately explore the impact of various prompt templates when applying LLM-as-a-Judge methods, leading to potentially inconsistent comparisons between different alignment algorithms. In this work, we systematically evaluate LLM-as-a-Judge on alignment tasks by defining more theoretically interpretable evaluation metrics and explicitly mitigating LLM internal inconsistency from reliability metrics. We develop an open-source framework to evaluate, compare, and visualize the reliability and alignment of LLM judges, which facilitates practitioners to choose LLM judges for alignment tasks. In the experiments, we examine effects of diverse prompt templates on LLM-judge reliability and also demonstrate our developed framework by comparing various LLM judges on two common alignment datasets (i.e., TL;DR Summarization and HH-RLHF-Helpfulness). Our results indicate a significant impact of prompt templates on LLM judge performance, as well as a mediocre alignment level between the tested LLM judges and human evaluators.

The rapid proliferation of benchmarks for evaluating large language models (LLMs) has created an urgent need for systematic methods to assess benchmark quality itself. We propose Benchmark^2, a comprehensive framework comprising three complementary metrics: (1) Cross-Benchmark Ranking Consistency, measuring whether a benchmark produces model rankings aligned with peer benchmarks; (2) Discriminability Score, quantifying a benchmark's ability to differentiate between models; and (3) Capability Alignment Deviation, identifying problematic instances where stronger models fail but weaker models succeed within the same model family. We conduct extensive experiments across 15 benchmarks spanning mathematics, reasoning, and knowledge domains, evaluating 11 LLMs across four model families. Our analysis reveals significant quality variations among existing benchmarks and demonstrates that selective benchmark construction based on our metrics can achieve comparable evaluation performance with substantially reduced test sets.

Large language models (LLMs) have achieved remarkable capabilities but remain vulnerable to adversarial prompts known as jailbreaks, which can bypass safety alignment and elicit harmful outputs. Despite growing efforts in LLM safety research, existing evaluations are often fragmented, focused on isolated attack or defense techniques, and lack systematic, reproducible analysis. In this work, we introduce PandaGuard, a unified and modular framework that models LLM jailbreak safety as a multi-agent system comprising attackers, defenders, and judges. Our framework implements 19 attack methods and 12 defense mechanisms, along with multiple judgment strategies, all within a flexible plugin architecture supporting diverse LLM interfaces, multiple interaction modes, and configuration-driven experimentation that enhances reproducibility and practical deployment. Built on this framework, we develop PandaBench, a comprehensive benchmark that evaluates the interactions between these attack/defense methods across 49 LLMs and various judgment approaches, requiring over 3 billion tokens to execute. Our extensive evaluation reveals key insights into model vulnerabilities, defense cost-performance trade-offs, and judge consistency. We find that no single defense is optimal across all dimensions and that judge disagreement introduces nontrivial variance in safety assessments. We release the code, configurations, and evaluation results to support transparent and reproducible research in LLM safety.

As large language models (LLMs) become increasingly common in educational applications, there is a growing need for evidence-based methods to design and evaluate LLM prompts that produce personalized and pedagogically aligned out-puts. This study presents a generalizable, systematic approach for evaluating prompts, demonstrated through an analysis of LLM-generated follow-up questions in a structured dialogue activity. Six prompt templates were designed and tested. The templates incorporated established prompt engineering patterns, with each prompt emphasizing distinct pedagogical strategies. The prompt templates were compared through a tournament-style evaluation framework that can be adapted for other educational applications. The tournament employed the Glicko2 rating system with eight judges evaluating question pairs across three dimensions: format, dialogue support, and appropriateness for learners. Data was sourced from 120 authentic user interactions across three distinct educational deployments. Results showed that a single prompt related to strategic reading out-performed other templates with win probabilities ranging from 81% to 100% in pairwise comparisons. This prompt combined persona and context manager pat-terns and was designed to support metacognitive learning strategies such as self-directed learning. The methodology showcases how educational technology re- searchers can systematically evaluate and improve prompt designs, moving beyond ad-hoc prompt engineering toward evidence-based prompt development for educational applications.

The rise of LLM-based agents has opened new frontiers in AI applications, yet evaluating these agents remains a complex and underdeveloped area. This survey provides an in-depth overview of the emerging field of LLM agent evaluation, introducing a two-dimensional taxonomy that organizes existing work along (1) evaluation objectives -- what to evaluate, such as agent behavior, capabilities, reliability, and safety -- and (2) evaluation process -- how to evaluate, including interaction modes, datasets and benchmarks, metric computation methods, and tooling. In addition to taxonomy, we highlight enterprise-specific challenges, such as role-based access to data, the need for reliability guarantees, dynamic and long-horizon interactions, and compliance, which are often overlooked in current research. We also identify future research directions, including holistic, more realistic, and scalable evaluation. This work aims to bring clarity to the fragmented landscape of agent evaluation and provide a framework for systematic assessment, enabling researchers and practitioners to evaluate LLM agents for real-world deployment.

The rapid advancement of Large Language Models (LLMs) has driven their expanding application across various fields. One of the most promising applications is their role as evaluators based on natural language responses, referred to as ''LLMs-as-judges''. This framework has attracted growing attention from both academia and industry due to their excellent effectiveness, ability to generalize across tasks, and interpretability in the form of natural language. This paper presents a comprehensive survey of the LLMs-as-judges paradigm from five key perspectives: Functionality, Methodology, Applications, Meta-evaluation, and Limitations. We begin by providing a systematic definition of LLMs-as-Judges and introduce their functionality (Why use LLM judges?). Then we address methodology to construct an evaluation system with LLMs (How to use LLM judges?). Additionally, we investigate the potential domains for their application (Where to use LLM judges?) and discuss methods for evaluating them in various contexts (How to evaluate LLM judges?). Finally, we provide a detailed analysis of the limitations of LLM judges and discuss potential future directions. Through a structured and comprehensive analysis, we aim aims to provide insights on the development and application of LLMs-as-judges in both research and practice. We will continue to maintain the relevant resource list at https://github.com/CSHaitao/Awesome-LLMs-as-Judges.

While the deployment of large language models (LLMs) in high-value industries continues to expand, the systematic assessment of their safety against jailbreak and prompt-based attacks remains insufficient. Existing safety evaluation benchmarks and frameworks are often limited by an imbalanced integration of core components (attack, defense, and evaluation methods) and an isolation between flexible evaluation frameworks and standardized benchmarking capabilities. These limitations hinder reliable cross-study comparisons and create unnecessary overhead for comprehensive risk assessment. To address these gaps, we present TeleAI-Safety, a modular and reproducible framework coupled with a systematic benchmark for rigorous LLM safety evaluation. Our framework integrates a broad collection of 19 attack methods (including one self-developed method), 29 defense methods, and 19 evaluation methods (including one self-developed method). With a curated attack corpus of 342 samples spanning 12 distinct risk categories, the TeleAI-Safety benchmark conducts extensive evaluations across 14 target models. The results reveal systematic vulnerabilities and model-specific failure cases, highlighting critical trade-offs between safety and utility, and identifying potential defense patterns for future optimization. In practical scenarios, TeleAI-Safety can be flexibly adjusted with customized attack, defense, and evaluation combinations to meet specific demands. We release our complete code and evaluation results to facilitate reproducible research and establish unified safety baselines.

Large Language Models (LLMs) are widely used as proxies for human labelers in both training (Reinforcement Learning from AI Feedback) and large-scale response evaluation (LLM-as-a-judge). Alignment and evaluation are critical components in the development of reliable LLMs, and the choice of feedback protocol plays a central role in both but remains understudied. In this work, we show that the choice of feedback protocol for evaluation (absolute scores versus relative preferences) can significantly affect evaluation reliability and induce systematic biases. In the context of LLM-as-a-judge evaluation, we show that pairwise protocols are more vulnerable to distracted evaluation. Generator models can exploit spurious attributes (or distractor features) favored by the LLM judge, resulting in inflated scores for lower-quality outputs. We find that absolute scoring is more robust to such manipulation, producing judgments that better reflect response quality and are less influenced by distractor features. Our results demonstrate that generator models can flip preferences by embedding distractor features, skewing LLM-as-a-judge comparisons and leading to inaccurate conclusions about model quality in benchmark evaluations. Pairwise preferences flip in about 35% of the cases, compared to only 9% for absolute scores. We offer recommendations for choosing feedback protocols based on dataset characteristics and evaluation objectives.

This paper introduces a comprehensive benchmark for evaluating how Large Language Models (LLMs) respond to linguistic shibboleths: subtle linguistic markers that can inadvertently reveal demographic attributes such as gender, social class, or regional background. Through carefully constructed interview simulations using 100 validated question-response pairs, we demonstrate how LLMs systematically penalize certain linguistic patterns, particularly hedging language, despite equivalent content quality. Our benchmark generates controlled linguistic variations that isolate specific phenomena while maintaining semantic equivalence, which enables the precise measurement of demographic bias in automated evaluation systems. We validate our approach along multiple linguistic dimensions, showing that hedged responses receive 25.6% lower ratings on average, and demonstrate the benchmark's effectiveness in identifying model-specific biases. This work establishes a foundational framework for detecting and measuring linguistic discrimination in AI systems, with broad applications to fairness in automated decision-making contexts.

As Large Language Models (LLMs) saturate elementary benchmarks, the research frontier has shifted from generation to the reliability of automated evaluation. We demonstrate that standard "LLM-as-a-Judge" protocols suffer from a systematic Alignment Gap when applied to upper-undergraduate to early graduate level mathematics. To quantify this, we introduce QEDBench, the first large-scale dual-rubric alignment benchmark to systematically measure alignment with human experts on university-level math proofs by contrasting course-specific rubrics against expert common knowledge criteria. By deploying a dual-evaluation matrix (7 judges x 5 solvers) against 1,000+ hours of human evaluation, we reveal that certain frontier evaluators like Claude Opus 4.5, DeepSeek-V3, Qwen 2.5 Max, and Llama 4 Maverick exhibit significant positive bias (up to +0.18, +0.20, +0.30, +0.36 mean score inflation, respectively). Furthermore, we uncover a critical reasoning gap in the discrete domain: while Gemini 3.0 Pro achieves state-of-the-art performance (0.91 average human evaluation score), other reasoning models like GPT-5 Pro and Claude Sonnet 4.5 see their performance significantly degrade in discrete domains. Specifically, their average human evaluation scores drop to 0.72 and 0.63 in Discrete Math, and to 0.74 and 0.50 in Graph Theory. In addition to these research results, we also release QEDBench as a public benchmark for evaluating and improving AI judges. Our benchmark is publicly published at https://github.com/qqliu/Yale-QEDBench.

Code-switching (CSW) is the act of alternating between two or more languages within a single discourse. This phenomenon is widespread in multilingual communities, and increasingly prevalent in online content, where users naturally mix languages in everyday communication. As a result, Large Language Models (LLMs), now central to content processing and generation, are frequently exposed to code-switched inputs. Given their widespread use, it is crucial to understand how LLMs process and reason about such mixed-language text. This paper presents a systematic evaluation of LLM comprehension under code-switching by generating CSW variants of established reasoning and comprehension benchmarks. While degradation is evident when foreign tokens disrupt English textx2013even under linguistic constraintsx2013embedding English into other languages often improves comprehension. Though prompting yields mixed results, fine-tuning offers a more stable path to degradation mitigation.

Code-switching is a pervasive phenomenon in multilingual communication, yet the robustness of large language models (LLMs) in mixed-language settings remains insufficiently understood. In this work, we present a comprehensive evaluation of LLM capabilities in understanding, reasoning over, and generating code-switched text. We introduce CodeMixQA a novel benchmark with high-quality human annotations, comprising 16 diverse parallel code-switched language-pair variants that span multiple geographic regions and code-switching patterns, and include both original scripts and their transliterated forms. Using this benchmark, we analyze the reasoning behavior of LLMs on code-switched question-answering tasks, shedding light on how models process and reason over mixed-language inputs. We further conduct a systematic evaluation of LLM-generated synthetic code-switched text, focusing on both naturalness and semantic fidelity, and uncover key limitations in current generation capabilities. Our findings reveal persistent challenges in both reasoning and generation under code-switching conditions and provide actionable insights for building more robust multilingual LLMs. We release the dataset and code as open source.

We present an open-source benchmark and evaluation framework for assessing emotional boundary handling in Large Language Models (LLMs). Using a dataset of 1156 prompts across six languages, we evaluated three leading LLMs (GPT-4o, Claude-3.5 Sonnet, and Mistral-large) on their ability to maintain appropriate emotional boundaries through pattern-matched response analysis. Our framework quantifies responses across seven key patterns: direct refusal, apology, explanation, deflection, acknowledgment, boundary setting, and emotional awareness. Results demonstrate significant variation in boundary-handling approaches, with Claude-3.5 achieving the highest overall score (8.69/10) and producing longer, more nuanced responses (86.51 words on average). We identified a substantial performance gap between English (average score 25.62) and non-English interactions (< 0.22), with English responses showing markedly higher refusal rates (43.20% vs. < 1% for non-English). Pattern analysis revealed model-specific strategies, such as Mistral's preference for deflection (4.2%) and consistently low empathy scores across all models (< 0.06). Limitations include potential oversimplification through pattern matching, lack of contextual understanding in response analysis, and binary classification of complex emotional responses. Future work should explore more nuanced scoring methods, expand language coverage, and investigate cultural variations in emotional boundary expectations. Our benchmark and methodology provide a foundation for systematic evaluation of LLM emotional intelligence and boundary-setting capabilities.

Large Language Models (LLMs) such as ChatGPT, Claude, and Gemini increasingly act as general-purpose copilots, yet they often respond with unnecessary length on simple requests, adding redundant explanations, hedging, or boilerplate that increases cognitive load and inflates token-based inference cost. Prior work suggests that preference-based post-training and LLM-judged evaluations can induce systematic length bias, where longer answers are rewarded even at comparable quality. We introduce YapBench, a lightweight benchmark for quantifying user-visible over-generation on brevity-ideal prompts. Each item consists of a single-turn prompt, a curated minimal-sufficient baseline answer, and a category label. Our primary metric, YapScore, measures excess response length beyond the baseline in characters, enabling comparisons across models without relying on any specific tokenizer. We summarize model performance via the YapIndex, a uniformly weighted average of category-level median YapScores. YapBench contains over three hundred English prompts spanning three common brevity-ideal settings: (A) minimal or ambiguous inputs where the ideal behavior is a short clarification, (B) closed-form factual questions with short stable answers, and (C) one-line coding tasks where a single command or snippet suffices. Evaluating 76 assistant LLMs, we observe an order-of-magnitude spread in median excess length and distinct category-specific failure modes, including vacuum-filling on ambiguous inputs and explanation or formatting overhead on one-line technical requests. We release the benchmark and maintain a live leaderboard for tracking verbosity behavior over time.

The recent surge in the capabilities of Large Language Models (LLMs) has led to claims that they are approaching a level of creativity akin to human capabilities. This idea has sparked a blend of excitement and apprehension. However, a critical piece that has been missing in this discourse is a systematic evaluation of LLM creativity, particularly in comparison to human divergent thinking. To bridge this gap, we leverage recent advances in creativity science to build a framework for in-depth analysis of divergent creativity in both state-of-the-art LLMs and a substantial dataset of 100,000 humans. We found evidence suggesting that LLMs can indeed surpass human capabilities in specific creative tasks such as divergent association and creative writing. Our quantitative benchmarking framework opens up new paths for the development of more creative LLMs, but it also encourages more granular inquiries into the distinctive elements that constitute human inventive thought processes, compared to those that can be artificially generated.

Classical and natural language planning tasks remain a difficult domain for modern large language models (LLMs). In this work, we lay the foundations for improving planning capabilities of LLMs. First, we construct a comprehensive benchmark suite encompassing both classical planning benchmarks and natural language scenarios. This suite includes algorithms to methodically generate instances of tasks with varying levels of difficulty, allowing for rigorous and systematic evaluation of LLM performance. Next, we investigate the use of many-shot in-context learning to enhance LLM planning, exploring the relationship between increased context length and improved planning performance. In addition, we demonstrate the positive impact of fine-tuning LLMs on optimal planning paths. We also probe the efficacy of chain-of-thought reasoning methods to improve LLM planning performance. Moreover, we probe the performance of the proposed methods in out-of-distribution scenarios, assessing the ability to generalize to novel and unseen planning challenges. Finally, we investigate model's failure modes and reveal insights that hold true across different benchmarks.

As large language models (LLMs) evolve from static chatbots into autonomous agents, the primary vulnerability surface shifts from final outputs to intermediate execution traces. While safety guardrails are well-benchmarked for natural language responses, their efficacy remains largely unexplored within multi-step tool-use trajectories. To address this gap, we introduce TraceSafe-Bench, the first comprehensive benchmark specifically designed to assess mid-trajectory safety. It encompasses 12 risk categories, ranging from security threats (e.g., prompt injection, privacy leaks) to operational failures (e.g., hallucinations, interface inconsistencies), featuring over 1,000 unique execution instances. Our evaluation of 13 LLM-as-a-guard models and 7 specialized guardrails yields three critical findings: 1) Structural Bottleneck: Guardrail efficacy is driven more by structural data competence (e.g., JSON parsing) than semantic safety alignment. Performance correlates strongly with structured-to-text benchmarks (ρ=0.79) but shows near-zero correlation with standard jailbreak robustness. 2) Architecture over Scale: Model architecture influences risk detection performance more significantly than model size, with general-purpose LLMs consistently outperforming specialized safety guardrails in trajectory analysis. 3) Temporal Stability: Accuracy remains resilient across extended trajectories. Increased execution steps allow models to pivot from static tool definitions to dynamic execution behaviors, actually improving risk detection performance in later stages. Our findings suggest that securing agentic workflows requires jointly optimizing for structural reasoning and safety alignment to effectively mitigate mid-trajectory risks.

Generating accurate step-by-step reasoning is essential for Large Language Models (LLMs) to address complex problems and enhance robustness and interpretability. Despite the flux of research on developing advanced reasoning approaches, systematically analyzing the diverse LLMs and reasoning strategies in generating reasoning chains remains a significant challenge. The difficulties stem from the lack of two key elements: (1) an automatic method for evaluating the generated reasoning chains on different tasks, and (2) a unified formalism and implementation of the diverse reasoning approaches for systematic comparison. This paper aims to close the gap: (1) We introduce AutoRace for fully automated reasoning chain evaluation. Existing metrics rely on expensive human annotations or pre-defined LLM prompts not adaptable to different tasks. In contrast, AutoRace automatically creates detailed evaluation criteria tailored for each task, and uses GPT-4 for accurate evaluation following the criteria. (2) We develop LLM Reasoners, a library for standardized modular implementation of existing and new reasoning algorithms, under a unified formulation of the search, reward, and world model components. With the new evaluation and library, (3) we conduct extensive study of different reasoning approaches (e.g., CoT, ToT, RAP). The analysis reveals interesting findings about different factors contributing to reasoning, including the reward-guidance, breadth-vs-depth in search, world model, and prompt formats, etc.

LLM-as-a-Judge has emerged as a promising alternative to human evaluators across various tasks, yet inherent biases - particularly position bias, the tendency to favor solutions based on their position within the prompt - compromise its reliability. This exploratory study evaluates position bias in LLM judges across pairwise and list-wise comparison settings, introducing three metrics: repetition stability, position consistency, and preference fairness. Our experiments, involving 15 LLM judges across MTBench and DevBench with 22 tasks and approximately 40 solution-generating models, result in over 150,000 evaluation instances. We identify Judge-Level, Candidate-Level, and Task-Level factors contributing to bias. The findings confirm that position bias is not due to random chance and varies significantly across judges and tasks. While position bias is weakly influenced by the length of prompt components, it is strongly affected by the quality gap between solutions. Our agreement and disagreement analysis among judges further provides insights into the distribution of judging difficulty across the dataset, and highlights the potential for dataset modifications.

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.

To meet the ever-increasing demand for computational efficiency, Neural Processing Units (NPUs) have become critical in modern AI infrastructure. However, unlocking their full potential requires developing high-performance compute kernels using vendor-specific Domain-Specific Languages (DSLs), a task that demands deep hardware expertise and is labor-intensive. While Large Language Models (LLMs) have shown promise in general code generation, they struggle with the strict constraints and scarcity of training data in the NPU domain. Our preliminary study reveals that state-of-the-art general-purpose LLMs fail to generate functional complex kernels for Ascend NPUs, yielding a near-zero success rate. To address these challenges, we propose AscendKernelGen, a generation-evaluation integrated framework for NPU kernel development. We introduce Ascend-CoT, a high-quality dataset incorporating chain-of-thought reasoning derived from real-world kernel implementations, and KernelGen-LM, a domain-adaptive model trained via supervised fine-tuning and reinforcement learning with execution feedback. Furthermore, we design NPUKernelBench, a comprehensive benchmark for assessing compilation, correctness, and performance across varying complexity levels. Experimental results demonstrate that our approach significantly bridges the gap between general LLMs and hardware-specific coding. Specifically, the compilation success rate on complex Level-2 kernels improves from 0% to 95.5% (Pass@10), while functional correctness achieves 64.3% compared to the baseline's complete failure. These results highlight the critical role of domain-specific reasoning and rigorous evaluation in automating accelerator-aware code generation.

Large language models (LLMs) have shown promise for automated patching, but their effectiveness depends strongly on how they are integrated into patching systems. While prior work explores prompting strategies and individual agent designs, the field lacks a systematic comparison of patching architectures. In this paper, we present a controlled evaluation of four LLM-based patching paradigms -- fixed workflow, single-agent system, multi-agent system, and general-purpose code agents -- using a unified benchmark and evaluation framework. We analyze patch correctness, failure modes, token usage, and execution time across real-world vulnerability tasks. Our results reveal clear architectural trade-offs: fixed workflows are efficient but brittle, single-agent systems balance flexibility and cost, and multi-agent designs improve generalization at the expense of substantially higher overhead and increased risk of reasoning drift on complex tasks. Surprisingly, general-purpose code agents achieve the strongest overall patching performance, benefiting from general-purpose tool interfaces that support effective adaptation across vulnerability types. Overall, we show that architectural design and iteration depth, rather than model capability alone, dominate the reliability and cost of LLM-based automated patching.

This research offers a unique evaluation of how AI systems interpret the digital language of Generation Alpha (Gen Alpha, born 2010-2024). As the first cohort raised alongside AI, Gen Alpha faces new forms of online risk due to immersive digital engagement and a growing mismatch between their evolving communication and existing safety tools. Their distinct language, shaped by gaming, memes, and AI-driven trends, often conceals harmful interactions from both human moderators and automated systems. We assess four leading AI models (GPT-4, Claude, Gemini, and Llama 3) on their ability to detect masked harassment and manipulation within Gen Alpha discourse. Using a dataset of 100 recent expressions from gaming platforms, social media, and video content, the study reveals critical comprehension failures with direct implications for online safety. This work contributes: (1) a first-of-its-kind dataset capturing Gen Alpha expressions; (2) a framework to improve AI moderation systems for youth protection; (3) a multi-perspective evaluation including AI systems, human moderators, and parents, with direct input from Gen Alpha co-researchers; and (4) an analysis of how linguistic divergence increases youth vulnerability. Findings highlight the urgent need to redesign safety systems attuned to youth communication, especially given Gen Alpha reluctance to seek help when adults fail to understand their digital world. This study combines the insight of a Gen Alpha researcher with systematic academic analysis to address critical digital safety challenges.

Large language models (LLMs) exhibit a unified "general factor" of capability across 10 benchmarks, a finding confirmed by our factor analysis of 156 models, yet they still struggle with simple, trivial tasks for humans. This is because current benchmarks focus on task completion, failing to probe the foundational cognitive abilities that highlight these behaviors. We address this by introducing the NeuroCognition benchmark, grounded in three adapted neuropsychological tests: Raven's Progressive Matrices (abstract relational reasoning), Spatial Working Memory (maintenance and systematic search), and the Wisconsin Card Sorting Test (cognitive flexibility). Our evaluation reveals that while models perform strongly on text, their performance degrades for images and with increased complexity. Furthermore, we observe that complex reasoning is not universally beneficial, whereas simple, human-like strategies yield partial gains. We also find that NeuroCognition correlates positively with standard general-capability benchmarks, while still measuring distinct cognitive abilities beyond them. Overall, NeuroCognition emphasizes where current LLMs align with human-like intelligence and where they lack core adaptive cognition, showing the potential to serve as a verifiable, scalable source for improving LLMs.

Evaluation frameworks for text summarization have evolved in terms of both domain coverage and metrics. However, existing benchmarks still lack domain-specific assessment criteria, remain predominantly English-centric, and face challenges with human annotation due to the complexity of reasoning. To address these, we introduce MSumBench, which provides a multi-dimensional, multi-domain evaluation of summarization in English and Chinese. It also incorporates specialized assessment criteria for each domain and leverages a multi-agent debate system to enhance annotation quality. By evaluating eight modern summarization models, we discover distinct performance patterns across domains and languages. We further examine large language models as summary evaluators, analyzing the correlation between their evaluation and summarization capabilities, and uncovering systematic bias in their assessment of self-generated summaries. Our benchmark dataset is publicly available at https://github.com/DISL-Lab/MSumBench.

As LLM-based Multi-Agent Systems (MAS) are increasingly deployed for complex tasks, ensuring their reliability has become a pressing challenge. Since MAS coordinate through unstructured natural language rather than rigid protocols, they are prone to semantic failures (e.g., hallucinations, misinterpreted instructions, and reasoning drift) that propagate silently without raising runtime exceptions. Prevailing evaluation approaches, which measure only end-to-end task success, offer limited insight into how these failures arise or how effectively agents recover from them. To bridge this gap, we propose MAS-FIRE, a systematic framework for fault injection and reliability evaluation of MAS. We define a taxonomy of 15 fault types covering intra-agent cognitive errors and inter-agent coordination failures, and inject them via three non-invasive mechanisms: prompt modification, response rewriting, and message routing manipulation. Applying MAS-FIRE to three representative MAS architectures, we uncover a rich set of fault-tolerant behaviors that we organize into four tiers: mechanism, rule, prompt, and reasoning. This tiered view enables fine-grained diagnosis of where and why systems succeed or fail. Our findings reveal that stronger foundation models do not uniformly improve robustness. We further show that architectural topology plays an equally decisive role, with iterative, closed-loop designs neutralizing over 40% of faults that cause catastrophic collapse in linear workflows. MAS-FIRE provides the process-level observability and actionable guidance needed to systematically improve multi-agent systems.

With the growing use of large language models(LLMs) as evaluators, their application has expanded to code evaluation tasks, where they assess the correctness of generated code without relying on reference implementations. While this offers scalability and flexibility, it also raises a critical, unresolved question: Can LLM judges fairly and robustly evaluate semantically equivalent code with superficial variations? Functionally correct code often exhibits variations-such as differences in variable names, comments, or formatting-that should not influence its correctness. Yet, whether LLM judges can reliably handle these variations remains unclear. We present the first comprehensive study of this issue, defining six types of potential bias in code evaluation and revealing their systematic impact on LLM judges. Across five programming languages and multiple LLMs, we empirically demonstrate that all tested LLM judges are susceptible to both positive and negative biases, resulting in inflated or unfairly low scores. Moreover, we observe that LLM judges remain vulnerable to these biases even when prompted to generate test cases before scoring, highlighting the need for more robust code evaluation methods.

The rapid advancement of Large Language Models (LLMs) has created new opportunities for Automated Penetration Testing (AutoPT), spawning numerous frameworks aimed at achieving end-to-end autonomous attacks. However, despite the proliferation of related studies, existing research generally lacks systematic architectural analysis and large-scale empirical comparisons under a unified benchmark. Therefore, this paper presents the first Systematization of Knowledge (SoK) focusing on the architectural design and comprehensive empirical evaluation of current LLM-based AutoPT frameworks. At systematization level, we comprehensively review existing framework designs across six dimensions: agent architecture, agent plan, agent memory, agent execution, external knowledge, and benchmarks. At empirical level, we conduct large-scale experiments on 13 representative open-source AutoPT frameworks and 2 baseline frameworks utilizing a unified benchmark. The experiments consumed over 10 billion tokens in total and generated more than 1,500 execution logs, which were manually reviewed and analyzed over four months by a panel of more than 15 researchers with expertise in cybersecurity. By investigating the latest progress in this rapidly developing field, we provide researchers with a structured taxonomy to understand existing LLM-based AutoPT frameworks and a large-scale empirical benchmark, along with promising directions for future research.

Multi-turn instruction following capability constitutes a core competency of large language models (LLMs) in real-world applications. Existing evaluation benchmarks predominantly focus on fine-grained constraint satisfaction and domain-specific capability assessment, yet overlook the crucial structural dependency between dialogue turns that distinguishes multi-turn from single-turn interactions. This structural dependency not only reflects user intent but also establishes a second dimension for instruction following evaluation beyond constraint satisfaction. To address this gap, we propose StructFlowBench, a multi-turn instruction following benchmark with structural flow modeling. The benchmark innovatively defines a structural flow framework comprising six fundamental inter-turn relationships, which not only introduces novel structural constraints for model evaluation but also serves as generation parameters for creating customized dialogue flows tailored to specific scenarios. Adopting established LLM-based automatic evaluation methodologies, we conduct systematic evaluations of 13 leading open-source and closed-source LLMs. Experimental results reveal significant deficiencies in current models' comprehension of multi-turn dialogue structures. The code is available at https://github.com/MLGroupJLU/StructFlowBench.

As AI assistants serve millions of users daily, evaluating user experience (UX) beyond general model capability has become increasingly important. We present UXBench, the first user-centric benchmark grounded in real user feedback signals for evaluating preference alignment and dialogue generation. The benchmark consists of three interconnected tasks, UX Judge, UX Eval, and UX Recovery, with 7,400 test instances extracted from over 70K interaction logs of a mainstream Chinese AI assistant. The dataset closely reflects real user distributions, covering 8 scenarios, 83 domains, and diverse failure patterns that pose severe challenges. Extensive experiments on 26 frontier language models provide novel insights into how well models perceive user experience and how improvements in model capability contribute to better dialogue engagement. Through comprehensive analysis of model behavior and performance gaps, we show that user feedback prediction is a learnable capability, where a reward model trained from in-the-wild feedback signals can achieve well-calibrated accuracy. We further document the systematic biases of LLM-as-a-judge evaluation protocols and compare typical response strategies that directly affect user experience. UXBench establishes a new evaluation landscape and calls for greater attention to tailored UX optimization, contributing to a user-centric scaling law that shapes the success of AI assistants.

Fact verification is essential for ensuring the reliability of LLM applications. In this study, we evaluate 12 pre-trained LLMs and one specialized fact-verifier, including frontier LLMs and open-weight reasoning LLMs, using a collection of examples from 14 fact-checking benchmarks. We share three findings intended to guide future development of more robust fact verifiers. First, we highlight the importance of addressing annotation errors and ambiguity in datasets, demonstrating that approximately 16\% of ambiguous or incorrectly labeled data substantially influences model rankings. Neglecting this issue may result in misleading conclusions during comparative evaluations, and we suggest using a systematic pipeline utilizing LLM-as-a-judge to help identify these issues at scale. Second, we discover that frontier LLMs with few-shot in-context examples, often overlooked in previous works, achieve top-tier performance. We therefore recommend future studies include comparisons with these simple yet highly effective baselines. Lastly, despite their effectiveness, frontier LLMs incur substantial costs, motivating the development of small, fine-tuned fact verifiers. We show that these small models still have room for improvement, particularly on instances that require complex reasoning. Encouragingly, we demonstrate that augmenting training with synthetic multi-hop reasoning data significantly enhances their capabilities in such instances. We release our code, model, and dataset at https://github.com/just1nseo/verifying-the-verifiers

Large Language Models (LLMs) are increasingly deployed as scientific AI as- sistants, and a growing body of benchmarks evaluates their capabilities across knowledge retrieval, reasoning, code generation, and tool use. These evaluations, however, typically assume the scientific problem is already well-posed, whereas practical scientific assistance often begins with an ill-posed user request that must be refined through dialogue before any computation, analysis, or experiment can be carried out reliably. We introduce SCICONVBENCH, a benchmark for multi- turn clarification in scientific task formulation across four computational science problem domains: fluid mechanics, solid mechanics, materials science, and par- tial differential equations (PDEs). SCICONVBENCH targets two complementary capabilities: eliciting missing information (disambiguation) and detecting and correcting erroneous requests containing internally contradictory information (in- consistency resolution). Our benchmark pairs a structured task ontology with a rubric-based evaluation framework, enabling systematic measurement of LLM per- formance across three dimensions: clarification behavior, conversational grounding, and final-specification fidelity. Current frontier models perform relatively well on inconsistency resolution, but even the best model resolves only 52.7% of the disambiguation cases in fluid mechanics. We further find that frontier LLMs fre- quently make silent assumptions and perform implicit specification repairs that are not grounded in the conversation with users. SCICONVBENCH establishes a foundation for evaluating the upstream conversational reasoning that a reliable computational science assistant requires. The code and data can be found at https://github.com/csml-rpi/SciConvBench.

Scientific innovation is undergoing a paradigm shift driven by the rapid advancement of Large Language Models (LLMs). As science faces mounting challenges including information overload, disciplinary silos, and diminishing returns on conventional research methods, LLMs are emerging as powerful agents capable not only of enhancing scientific workflows but also of participating in and potentially leading the innovation process. Existing surveys mainly focus on different perspectives, phrases, and tasks in scientific research and discovery, while they have limitations in understanding the transformative potential and role differentiation of LLM. This survey proposes a comprehensive framework to categorize the evolving roles of LLMs in scientific innovation across three hierarchical levels: Evaluator, Collaborator, and Scientist. We distinguish between LLMs' contributions to structured scientific research processes and open-ended scientific discovery, thereby offering a unified taxonomy that clarifies capability boundaries, evaluation criteria, and human-AI interaction patterns at each level. Through an extensive analysis of current methodologies, benchmarks, systems, and evaluation metrics, this survey delivers an in-depth and systematic synthesis on LLM-driven scientific innovation. We present LLMs not only as tools for automating existing processes, but also as catalysts capable of reshaping the epistemological foundations of science itself. This survey offers conceptual clarity, practical guidance, and theoretical foundations for future research, while also highlighting open challenges and ethical considerations in the pursuit of increasingly autonomous AI-driven science. Resources related to this survey can be accessed on GitHub at: https://github.com/haoxuan-unt2024/llm4innovation.

The use of large language models (LLMs) for automated code generation has emerged as a significant focus within AI research. As these pretrained models continue to evolve, their ability to understand and generate complex code structures has opened new possibilities for automating intricate programming tasks for the sake of accurate code generation. Although contemporary foundational models demonstrate promoting results, researchers continue to explore optimal post-training strategies to enhance code quality. These include supervised fine-tuning, retrieval-augmented generation (RAG), debugging, and many others. In this paper, we combine two widely used approaches namely multi-agent collaboration and runtime execution information-based debugging, for improving code generation functionality, reliability, and practical applicability. We perform an empirical study in order to extend the evaluation of the individual strategies as well as the proposed composition of the activities of both strategies. Our study use 19 LLMs to examines the performance of individual and the proposed strategies, offering comprehensive insights into how different programming activities compositions and training paradigms influence code generation effectiveness. In particular, we implement a chained system that combines both strategies to assess their combined impact on functional accuracy, code reliability, and generation latency using two benchmark datasets commonly used for code generation. Our findings provide valuable insights for organizations seeking robust AI-driven coding solutions by guiding them in selecting models that can better adapt to complex post-training strategies, ultimately fostering the adoption of more effective and reliable code generation technologies.

Coding agents and LLM-powered applications routinely send potentially sensitive content to cloud LLM APIs where it may be logged, retained, used for training, or subpoenaed. Existing privacy tooling focuses on network-level encryption and organization-level DLP, neither of which addresses the content of prompts themselves. We present a systematic empirical evaluation of eight techniques for privacy-preserving LLM requests: (A) local-only inference, (B) redaction with placeholder restoration, (C) semantic rephrasing, (D) Trusted Execution Environment hosted inference, (E) split inference, (F) fully homomorphic encryption, (G) secret sharing via multi-party computation, and (H) differential-privacy noise. We implement all eight (or a tractable research-stage subset where deployment is not yet feasible) in an open-source shim compatible with MCP and any OpenAI-compatible API. We evaluate the four practical options (A, B, C, H) and their combinations across four workload classes using a ground-truth-labelled leak benchmark of 1,300 samples with 4,014 annotations. Our headline finding is that no single technique dominates: the combination A+B+C (route locally when possible, redact and rephrase the rest) achieves 0.6% combined leak on PII and 31.3% on proprietary code, with zero exact leaks on PII across 500 samples. We present a decision rule that selects the appropriate option(s) from a threat-model budget and workload characterisation. Code, benchmarks, and evaluation harness are released at https://github.com/jayluxferro/llm-redactor.

In enterprise settings, efficiently retrieving relevant information from large and complex knowledge bases is essential for operational productivity and informed decision-making. This research presents a systematic framework for metadata enrichment using large language models (LLMs) to enhance document retrieval in Retrieval-Augmented Generation (RAG) systems. Our approach employs a comprehensive, structured pipeline that dynamically generates meaningful metadata for document segments, substantially improving their semantic representations and retrieval accuracy. Through extensive experiments, we compare three chunking strategies-semantic, recursive, and naive-and evaluate their effectiveness when combined with advanced embedding techniques. The results demonstrate that metadata-enriched approaches consistently outperform content-only baselines, with recursive chunking paired with TF-IDF weighted embeddings yielding an 82.5% precision rate compared to 73.3% for semantic content-only approaches. The naive chunking strategy with prefix-fusion achieved the highest Hit Rate@10 of 0.925. Our evaluation employs cross-encoder reranking for ground truth generation, enabling rigorous assessment via Hit Rate and Metadata Consistency metrics. These findings confirm that metadata enrichment enhances vector clustering quality while reducing retrieval latency, making it a key optimization for RAG systems across knowledge domains. This work offers practical insights for deploying high-performance, scalable document retrieval solutions in enterprise settings, demonstrating that metadata enrichment is a powerful approach for enhancing RAG effectiveness.

Large language models (LLMs) have demonstrated excellent capabilities in generating structured diagrams from natural language descriptions. In particular, they have shown great promise in generating sequence diagrams for software engineering, typically represented in a text-based syntax such as Mermaid. However, systematic evaluations in this space remain underdeveloped as there is a lack of existing benchmarks to assess the LLM's correctness in this task. To address this shortcoming, we introduce MermaidSeqBench, a human-verified and LLM-synthetically-extended benchmark for assessing an LLM's capabilities in generating Mermaid sequence diagrams from textual prompts. The benchmark consists of a core set of 132 samples, starting from a small set of manually crafted and verified flows. These were expanded via a hybrid methodology combining human annotation, in-context LLM prompting, and rule-based variation generation. Our benchmark uses an LLM-as-a-judge model to assess Mermaid sequence diagram generation across fine-grained metrics, including syntax correctness, activation handling, error handling, and practical usability. We perform initial evaluations on numerous state-of-the-art LLMs and utilize multiple LLM judge models to demonstrate the effectiveness and flexibility of our benchmark. Our results reveal significant capability gaps across models and evaluation modes. Our proposed benchmark provides a foundation for advancing research in structured diagram generation and for developing more rigorous, fine-grained evaluation methodologies.

Parliamentary speech generation presents specific challenges for large language models beyond standard text generation tasks. Unlike general text generation, parliamentary speeches require not only linguistic quality but also political authenticity and ideological consistency. Current language models lack specialized training for parliamentary contexts, and existing evaluation methods focus on standard NLP metrics rather than political authenticity. To address this, we present ParliaBench, a benchmark for parliamentary speech generation. We constructed a dataset of speeches from UK Parliament to enable systematic model training. We introduce an evaluation framework combining computational metrics with LLM-as-a-judge assessments for measuring generation quality across three dimensions: linguistic quality, semantic coherence, and political authenticity. We propose two novel embedding-based metrics, Political Spectrum Alignment and Party Alignment, to quantify ideological positioning. We fine-tuned five large language models (LLMs), generated 28k speeches, and evaluated them using our framework, comparing baseline and fine-tuned models. Results show that fine-tuning produces statistically significant improvements across the majority of metrics and our novel metrics demonstrate strong discriminative power for political dimensions.

Large language models (LLMs) produce outputs with varying levels of uncertainty, and, just as often, varying levels of correctness; making their practical reliability far from guaranteed. To quantify this uncertainty, we systematically evaluate four approaches for confidence estimation in LLM outputs: VCE, MSP, Sample Consistency, and CoCoA (Vashurin et al., 2025). For the evaluation of the approaches, we conduct experiments on four question-answering tasks using a state-of-the-art open-source LLM. Our results show that each uncertainty metric captures a different facet of model confidence and that the hybrid CoCoA approach yields the best reliability overall, improving both calibration and discrimination of correct answers. We discuss the trade-offs of each method and provide recommendations for selecting uncertainty measures in LLM applications.

Large Language Models (LLMs) have achieved impressive results in Machine Translation (MT). However, careful evaluations by human reveal that the translations produced by LLMs still contain multiple errors. Importantly, feeding back such error information into the LLMs can lead to self-refinement and result in improved translation performance. Motivated by these insights, we introduce a systematic LLM-based self-refinement translation framework, named TEaR, which stands for Translate, Estimate, and Refine, marking a significant step forward in this direction. Our findings demonstrate that 1) our self-refinement framework successfully assists LLMs in improving their translation quality across a wide range of languages, whether it's from high-resource languages to low-resource ones or whether it's English-centric or centered around other languages; 2) TEaR exhibits superior systematicity and interpretability; 3) different estimation strategies yield varied impacts, directly affecting the effectiveness of the final corrections. Additionally, traditional neural translation models and evaluation models operate separately, often focusing on singular tasks due to their limited capabilities, while general-purpose LLMs possess the capability to undertake both tasks simultaneously. We further conduct cross-model correction experiments to investigate the potential relationship between the translation and evaluation capabilities of general-purpose LLMs. Our code and data are available at https://github.com/fzp0424/self_correct_mt

Feature engineering remains a cornerstone of tabular data analysis, and Large Language Models (LLMs) have emerged as a promising paradigm for its automation, giving rise to LLM-powered Automated Tabular Feature Engineering (LATTE). However, the field lacks standardized, cost-aware evaluation platforms, and the combinatorial explosion of design choices obscures true algorithmic progress. To bridge these gaps, we systematically deconstruct 15 representative LATTE methods into a unified 6-dimensional taxonomy. Based on this abstraction, we introduce LATTEArena, a standardized, modular, and extensible benchmarking framework that decouples monolithic pipelines into reusable execution blocks. By distilling the massive combinatorial space, we evaluate 24 core LATTE configurations across 7 research questions. Our head-to-head benchmarking goes beyond predictive accuracy to quantify token efficiency and execution robustness, yielding 17 empirical findings on cost-effectiveness trade-offs. Furthermore, we provide 3 concrete recommendations for optimal real-world deployment. By enabling controlled component-level comparisons, LATTEArena shifts the paradigm from ad-hoc prompt engineering to systematic context management. All code, datasets, and over 4,000 execution logs are publicly available to foster a dynamic, community-driven benchmark. Our framework, leaderboard, and all artifacts are hosted on the LATTEArena project website at https://goodenhak.github.io/LATTEArena.

Recent advances in large language models (LLMs) have enabled automatic generation of chain-of-thought (CoT) reasoning, leading to strong performance on tasks such as math and code. However, when reasoning steps reflect social stereotypes (e.g., those related to gender, race or age), they can reinforce harmful associations and lead to misleading conclusions. We present the first systematic evaluation of social bias within LLM-generated reasoning, using the BBQ dataset to analyze both prediction accuracy and bias. Our study spans a wide range of mainstream reasoning models, including instruction-tuned and CoT-augmented variants of DeepSeek-R1 (8B/32B), ChatGPT, and other open-source LLMs. We quantify how biased reasoning steps correlate with incorrect predictions and often lead to stereotype expression. To mitigate reasoning-induced bias, we propose Answer Distribution as Bias Proxy (ADBP), a lightweight mitigation method that detects bias by tracking how model predictions change across incremental reasoning steps. ADBP outperforms a stereotype-free baseline in most cases, mitigating bias and improving the accuracy of LLM outputs. Code will be released upon paper acceptance.

Large language model (LLM) post-training enhances latent skills, unlocks value alignment, improves performance, and enables domain adaptation. Unfortunately, post-training is known to induce forgetting, especially in the ubiquitous use-case of leveraging third-party pre-trained models, which is typically understood as a loss of parametric or factual knowledge. We argue that this accuracy-centric view is insufficient for modern foundation models and instead define forgetting as systematic model drift that degrades behavior and user experience. In this context, we introduce CapTrack, a capability-centric framework for analyzing forgetting in LLMs that combines a behavioral taxonomy with an evaluation suite built on established benchmarks and targeted adaptations. Using CapTrack, we conduct a large-scale empirical study across post-training algorithms, domains, and model families, including models up to 80B parameters. We find that forgetting extends beyond parametric knowledge, with pronounced drift in robustness and default behaviors. Instruction fine-tuning induces the strongest relative drift, while preference optimization is more conservative and can partially recover lost capabilities. Differences across model families persist, and no universal mitigation emerges.

The advent of Large Language Models (LLMs) has significantly revolutionized web search. The emergence of LLM-based Search Agents marks a pivotal shift towards deeper, dynamic, autonomous information seeking. These agents can comprehend user intentions and environmental context and execute multi-turn retrieval with dynamic planning, extending search capabilities far beyond the web. Leading examples like OpenAI's Deep Research highlight their potential for deep information mining and real-world applications. This survey provides the first systematic analysis of search agents. We comprehensively analyze and categorize existing works from the perspectives of architecture, optimization, application, and evaluation, ultimately identifying critical open challenges and outlining promising future research directions in this rapidly evolving field. Our repository is available on https://github.com/YunjiaXi/Awesome-Search-Agent-Papers.

Recent work on chain-of-thought (CoT) faithfulness reports single aggregate numbers (e.g., DeepSeek-R1 acknowledges hints 39% of the time), implying that faithfulness is an objective, measurable property of a model. This paper demonstrates that it is not. Three classifiers (a regex-only detector, a two-stage regex-plus-LLM pipeline, and an independent Claude Sonnet 4 judge) are applied to 10,276 influenced reasoning traces from 12 open-weight models spanning 9 families and 7B to 1T parameters. On identical data, these classifiers produce overall faithfulness rates of 74.4%, 82.6%, and 69.7%, respectively, with non-overlapping 95% confidence intervals. Per-model gaps range from 2.6 to 30.6 percentage points; all are statistically significant (McNemar's test, p < 0.001). The disagreements are systematic, not random: inter-classifier agreement measured by Cohen's kappa ranges from 0.06 ("slight") for sycophancy hints to 0.42 ("moderate") for grader hints, and the asymmetry is pronounced: for sycophancy, 883 cases are classified as faithful by the pipeline but unfaithful by the Sonnet judge, while only 2 go the other direction. Classifier choice can also reverse model rankings: Qwen3.5-27B ranks 1st under the pipeline but 7th under the Sonnet judge; OLMo-3.1-32B moves in the opposite direction, from 9th to 3rd. The root cause is that different classifiers operationalize related faithfulness constructs at different levels of stringency (lexical mention versus epistemic dependence), and these constructs yield divergent measurements on the same behavior. These results demonstrate that published faithfulness numbers cannot be meaningfully compared across studies that use different classifiers, and that future evaluations should report sensitivity ranges across multiple classification methodologies rather than single point estimates.

Large Language Models (LLMs) offer transformative capabilities for hardware design automation, particularly in Verilog code generation. However, they also pose significant data security challenges, including Verilog evaluation data contamination, intellectual property (IP) design leakage, and the risk of malicious Verilog generation. We introduce SALAD, a comprehensive assessment that leverages machine unlearning to mitigate these threats. Our approach enables the selective removal of contaminated benchmarks, sensitive IP and design artifacts, or malicious code patterns from pre-trained LLMs, all without requiring full retraining. Through detailed case studies, we demonstrate how machine unlearning techniques effectively reduce data security risks in LLM-aided hardware design.

As LLMs are increasingly used as judges in code applications, they should be evaluated in realistic interactive settings that capture partial context and ambiguous intent. We present TRACE (Tool for Rubric Analysis in Code Evaluation), a framework that evaluates LLM judges' ability to predict human preferences and automatically extracts rubric items to reveal systematic biases in how humans and models weigh each item. Across three modalities -- chat-based programming, IDE autocompletion, and instructed code editing -- we use TRACE to measure how well LLM judges align with developer preferences. Among 13 different models, the best judges underperform human annotators by 12-23%. TRACE identifies 35 significant sources of misalignment between humans and judges across interaction modalities, the majority of which correspond to existing software engineering code quality criteria. For example, in chat-based coding, judges are biased towards longer code explanations while humans prefer shorter ones. We find significant misalignment on the majority of existing code quality dimensions, showing alignment gaps between LLM judges and human preference in realistic coding applications.

The remarkable success of large language models (LLMs) has motivated researchers to adapt them as universal predictors for various graph tasks. As a widely recognized paradigm, Graph-Tokenizing LLMs (GTokenLLMs) compress complex graph data into graph tokens and treat them as prefix tokens for querying LLMs, leading many to believe that LLMs can understand graphs more effectively and efficiently. In this paper, we challenge this belief: Do GTokenLLMs fully understand graph tokens in the natural-language embedding space? Motivated by this question, we formalize a unified framework for GTokenLLMs and propose an evaluation pipeline, GTEval, to assess graph-token understanding via instruction transformations at the format and content levels. We conduct extensive experiments on 6 representative GTokenLLMs with GTEval. The primary findings are as follows: (1) Existing GTokenLLMs do not fully understand graph tokens. They exhibit over-sensitivity or over-insensitivity to instruction changes, and rely heavily on text for reasoning; (2) Although graph tokens preserve task-relevant graph information and receive attention across LLM layers, their utilization varies across models and instruction variants; (3) Additional instruction tuning can improve performance on the original and seen instructions, but it does not fully address the challenge of graph-token understanding, calling for further improvement.

Peer review is fundamental to scientific research, but the growing volume of publications has intensified the challenges of this expertise-intensive process. While LLMs show promise in various scientific tasks, their potential to assist with peer review, particularly in identifying paper limitations, remains understudied. We first present a comprehensive taxonomy of limitation types in scientific research, with a focus on AI. Guided by this taxonomy, for studying limitations, we present LimitGen, the first comprehensive benchmark for evaluating LLMs' capability to support early-stage feedback and complement human peer review. Our benchmark consists of two subsets: LimitGen-Syn, a synthetic dataset carefully created through controlled perturbations of high-quality papers, and LimitGen-Human, a collection of real human-written limitations. To improve the ability of LLM systems to identify limitations, we augment them with literature retrieval, which is essential for grounding identifying limitations in prior scientific findings. Our approach enhances the capabilities of LLM systems to generate limitations in research papers, enabling them to provide more concrete and constructive feedback.

LLM-as-a-Judge has been widely adopted across various research and practical applications, yet the robustness and reliability of its evaluation remain a critical issue. A core challenge it faces is bias, which has primarily been studied in terms of known biases and their impact on evaluation outcomes, while automated and systematic exploration of potential unknown biases is still lacking. Nevertheless, such exploration is crucial for enhancing the robustness and reliability of evaluations. To bridge this gap, we propose BiasScope, a LLM-driven framework for automatically and at scale discovering potential biases that may arise during model evaluation. BiasScope can uncover potential biases across different model families and scales, with its generality and effectiveness validated on the JudgeBench dataset. It overcomes the limitations of existing approaches, transforming bias discovery from a passive process relying on manual effort and predefined bias lists into an active and comprehensive automated exploration. Moreover, based on BiasScope, we propose JudgeBench-Pro, an extended version of JudgeBench and a more challenging benchmark for evaluating the robustness of LLM-as-a-judge. Strikingly, even powerful LLMs as evaluators show error rates above 50\% on JudgeBench-Pro, underscoring the urgent need to strengthen evaluation robustness and to mitigate potential biases further.

The rapid progression of multimodal large language models (MLLMs) has demonstrated superior performance on various multimodal benchmarks. However, the issue of data contamination during training creates challenges in performance evaluation and comparison. While numerous methods exist for detecting dataset contamination in large language models (LLMs), they are less effective for MLLMs due to their various modalities and multiple training phases. In this study, we introduce a multimodal data contamination detection framework, MM-Detect, designed for MLLMs. Our experimental results indicate that MM-Detect is sensitive to varying degrees of contamination and can highlight significant performance improvements due to leakage of the training set of multimodal benchmarks. Furthermore, We also explore the possibility of contamination originating from the pre-training phase of LLMs used by MLLMs and the fine-tuning phase of MLLMs, offering new insights into the stages at which contamination may be introduced.

The rapid development of Large Language Models (LLMs) has led to great strides in model capabilities like reasoning and long-context understanding. However, as LLMs are able to process longer contexts, it becomes more challenging to evaluate whether they have acquired certain capabilities, since the length of text (e.g., 100K tokens) they can process far exceeds what humans can reliably assess in a reasonable duration. In this paper, we propose using complex synthetic tasks as a proxy evaluation method, and present S3Eval, a Synthetic, Scalable, Systematic evaluation suite for LLMs evaluation. As a synthetic benchmark, S3Eval enables the creation of any number of evaluation examples that are theoretically invisible to LLMs, mitigating the test set contamination issue. The synthetic nature of S3Eval provides users full control over the dataset, allowing them to systematically probe LLM capabilities by scaling text length and varying task difficulty across diverse scenarios. The strong correlation between S3Eval performance and scores of real-world benchmarks like Big-Bench Hard (BBH) demonstrates the soundness of using S3Eval for evaluation of LLMs. The in-depth analysis also uncover additional insights, including performance drop when the answer is sparsely distributed or located in the middle context, as well as some counter-intuitive trends of model performance.

Large language models (LLMs) have demonstrated exceptional performance in text generation within current NLP research. However, the lack of factual accuracy is still a dark cloud hanging over the LLM skyscraper. Structural knowledge prompting (SKP) is a prominent paradigm to integrate external knowledge into LLMs by incorporating structural representations, achieving state-of-the-art results in many knowledge-intensive tasks. However, existing methods often focus on specific problems, lacking a comprehensive exploration of the generalization and capability boundaries of SKP. This paper aims to evaluate and rethink the generalization capability of the SKP paradigm from four perspectives including Granularity, Transferability, Scalability, and Universality. To provide a thorough evaluation, we introduce a novel multi-granular, multi-level benchmark called SUBARU, consisting of 9 different tasks with varying levels of granularity and difficulty.

We introduce TestCase-Eval, a new benchmark for systematic evaluation of LLMs in test-case generation. TestCase-Eval includes 500 algorithm problems and 100,000 human-crafted solutions from the Codeforces platform. It focuses on two pivotal tasks: (1) Fault Coverage, which measures how well LLM-generated test sets probe diverse input scenarios and cover a wide range of potential failure modes. (2) Fault Exposure, which evaluates whether LLMs can craft a tailored test input that reveals a specific incorrect code implementation. We provide a comprehensive assessment of 19 state-of-the-art open-source and proprietary LLMs on TestCase-Eval, offering insights into their strengths and limitations in generating effective test cases for algorithm problems.

Large language models (LLMs) are transforming the landscape of medicine, yet two fundamental challenges persist: keeping up with rapidly evolving medical knowledge and providing verifiable, evidence-grounded reasoning. Retrieval-augmented generation (RAG) has been widely adopted to address these limitations by supplementing model outputs with retrieved evidence. However, whether RAG reliably achieves these goals remains unclear. Here, we present the most comprehensive expert evaluation of RAG in medicine to date. Eighteen medical experts contributed a total of 80,502 annotations, assessing 800 model outputs generated by GPT-4o and Llama-3.1-8B across 200 real-world patient and USMLE-style queries. We systematically decomposed the RAG pipeline into three components: (i) evidence retrieval (relevance of retrieved passages), (ii) evidence selection (accuracy of evidence usage), and (iii) response generation (factuality and completeness of outputs). Contrary to expectation, standard RAG often degraded performance: only 22% of top-16 passages were relevant, evidence selection remained weak (precision 41-43%, recall 27-49%), and factuality and completeness dropped by up to 6% and 5%, respectively, compared with non-RAG variants. Retrieval and evidence selection remain key failure points for the model, contributing to the overall performance drop. We further show that simple yet effective strategies, including evidence filtering and query reformulation, substantially mitigate these issues, improving performance on MedMCQA and MedXpertQA by up to 12% and 8.2%, respectively. These findings call for re-examining RAG's role in medicine and highlight the importance of stage-aware evaluation and deliberate system design for reliable medical LLM applications.

Low-Resource Languages (LRLs) present significant challenges in natural language processing due to their limited linguistic resources and underrepresentation in standard datasets. While recent advancements in Large Language Models (LLMs) and Neural Machine Translation (NMT) have substantially improved translation capabilities for high-resource languages, performance disparities persist for LRLs, particularly impacting privacy-sensitive and resource-constrained scenarios. This paper systematically evaluates the limitations of current LLMs across 200 languages using benchmarks such as FLORES-200. We also explore alternative data sources, including news articles and bilingual dictionaries, and demonstrate how knowledge distillation from large pre-trained models can significantly improve smaller LRL translations. Additionally, we investigate various fine-tuning strategies, revealing that incremental enhancements markedly reduce performance gaps on smaller LLMs.

Large Language Model (LLM) agents are increasingly expected to maintain coherent, long-term personalized memory, yet current benchmarks primarily measure static fact retrieval, overlooking the ability to revise stored beliefs when new evidence emerges. We identify a critical and underexplored failure mode, Implicit Conflict: a later observation invalidates an earlier memory without explicit negation, requiring contextual inference and commonsense reasoning to detect. To rigorously evaluate this capability, we introduce STALE, a benchmark of 400 expert-validated conflict scenarios (1,200 evaluation queries across three probing dimensions) spanning over 100 everyday topics with contexts up to 150K tokens. We propose a three-dimensional probing framework that tests State Resolution (detecting that a prior belief is outdated), Premise Resistance (rejecting queries that falsely presuppose a stale state), and Implicit Policy Adaptation (proactively applying updated states in downstream behavior). A systematic evaluation of frontier LLMs and specialized memory frameworks reveals a pervasive gap between retrieving updated evidence and acting on it, with even the best evaluated model achieving only 55.2% overall accuracy. Models often accept outdated assumptions embedded in a user's query, and they struggle to recognize when a change in one aspect of the user's state should invalidate related memories. To establish an initial baseline for state-aware memory, we further present CUPMem, a prototype that strengthens write-time revision through structured state consolidation and propagation-aware search, suggesting that explicit state adjudication is a promising direction for robust agentic memory.

This paper presents a comprehensive evaluation framework for aligning Persian Large Language Models (LLMs) with critical ethical dimensions, including safety, fairness, and social norms. It addresses the gaps in existing LLM evaluation frameworks by adapting them to Persian linguistic and cultural contexts. This benchmark creates three types of Persian-language benchmarks: (i) translated data, (ii) new data generated synthetically, and (iii) new naturally collected data. We translate Anthropic Red Teaming data, AdvBench, HarmBench, and DecodingTrust into Persian. Furthermore, we create ProhibiBench-fa, SafeBench-fa, FairBench-fa, and SocialBench-fa as new datasets to address harmful and prohibited content in indigenous culture. Moreover, we collect extensive dataset as GuardBench-fa to consider Persian cultural norms. By combining these datasets, our work establishes a unified framework for evaluating Persian LLMs, offering a new approach to culturally grounded alignment evaluation. A systematic evaluation of Persian LLMs is performed across the three alignment aspects: safety (avoiding harmful content), fairness (mitigating biases), and social norms (adhering to culturally accepted behaviors). We present a publicly available leaderboard that benchmarks Persian LLMs with respect to safety, fairness, and social norms at: https://huggingface.co/spaces/MCILAB/LLM_Alignment_Evaluation.

Automated Code Review (ACR) is crucial for software quality, yet existing benchmarks often fail to reflect real-world complexities, hindering the evaluation of modern Large Language Models (LLMs). Current benchmarks frequently focus on fine-grained code units, lack complete project context, and use inadequate evaluation metrics. To address these limitations, we introduce SWRBench , a new benchmark comprising 1000 manually verified Pull Requests (PRs) from GitHub, offering PR-centric review with full project context. SWRBench employs an objective LLM-based evaluation method that aligns strongly with human judgment (~90 agreement) by verifying if issues from a structured ground truth are covered in generated reviews. Our systematic evaluation of mainstream ACR tools and LLMs on SWRBench reveals that current systems underperform, and ACR tools are more adept at detecting functional errors. Subsequently, we propose and validate a simple multi-review aggregation strategy that significantly boosts ACR performance, increasing F1 scores by up to 43.67%. Our contributions include the SWRBench benchmark, its objective evaluation method, a comprehensive study of current ACR capabilities, and an effective enhancement approach, offering valuable insights for advancing ACR research.

As the large language model (LLM) ecosystem expands, individual models exhibit varying capabilities across queries, benchmarks, and domains, motivating the development of LLM routing. While prior work has largely focused on router mechanism design, LLM profiles, which capture model capabilities, remain underexplored. In this work, we ask: How does LLM profile design affect routing performance across different routers? Addressing this question helps clarify the role of profiles in routing, disentangle profile design from router design, and enable fairer comparison and more principled development of routing systems. To this end, we view LLM profiling as a structured information integration problem over heterogeneous interaction histories. We develop a general design space of LLM profiles, named RouteProfile, along four key dimensions: organizational form, representation type, aggregation depth, and learning configuration. Through systematic evaluation across three representative routers under both standard and new-LLM generalization settings, we show that: (1) structured profiles consistently outperform flat ones; (2) query-level signals are more reliable than coarse domain-level signals; and (3) generalization to newly introduced models benefits most from structured profiles under trainable configurations. Overall, our work highlights LLM profile design as an important direction for future routing research.

Large language model based agents often fail in unfamiliar environments due to premature exploitation: a tendency to act on prior knowledge before acquiring sufficient environment-specific information. We identify autonomous exploration as a critical yet underexplored capability for building adaptive agents. To formalize and quantify this capability, we introduce Exploration Checkpoint Coverage, a verifiable metric that measures how broadly an agent discovers key states, objects, and affordances. Our systematic evaluation reveals that agents trained with standard task-oriented reinforcement learning consistently exhibit narrow and repetitive behaviors that impede downstream performance. To address this limitation, we develop a training strategy that interleaves task-execution rollouts and exploration rollouts, with each type of rollout optimized by its corresponding verifiable reward. Building on this training strategy, we propose the Explore-then-Act paradigm, which decouples information-gathering from task execution: agents first utilize an interaction budget to acquire grounded environmental knowledge, then leverage it for task resolution. Our results demonstrate that learning to systematically explore is imperative for building generalizable and real-world-ready agents.

Low-Rank Adaptation (LoRA) is the prevailing approach for efficient large language model (LLM) fine-tuning. Building on this paradigm, recent studies have proposed alternative initialization strategies and architectural modifications, reporting substantial improvements over vanilla LoRA. However, these gains are often demonstrated under fixed or narrowly tuned hyperparameter settings, despite the known sensitivity of neural networks to training configurations. In this work, we systematically re-evaluate four representative LoRA variants alongside vanilla LoRA through extensive hyperparameter searches. Across mathematical and code generation tasks on diverse model scales, we find that different LoRA methods favor distinct learning rate ranges. Crucially, once learning rates are properly tuned, all methods achieve similar peak performance (within 1-2%), with only subtle rank-dependent behaviors. These results suggest that vanilla LoRA remains a competitive baseline and that improvements reported under single training configuration may not reflect consistent methodological advantages. Finally, a second-order analysis attributes the differing optimal learning rate ranges to variations in the largest Hessian eigenvalue, aligning with classical learning theories.

As large language models (LLMs) evolve into sophisticated autonomous agents capable of complex software development tasks, evaluating their real-world capabilities becomes critical. While existing benchmarks like LoCoBench~qiu2025locobench assess long-context code understanding, they focus on single-turn evaluation and cannot capture the multi-turn interactive nature, tool usage patterns, and adaptive reasoning required by real-world coding agents. We introduce LoCoBench-Agent, a comprehensive evaluation framework specifically designed to assess LLM agents in realistic, long-context software engineering workflows. Our framework extends LoCoBench's 8,000 scenarios into interactive agent environments, enabling systematic evaluation of multi-turn conversations, tool usage efficiency, error recovery, and architectural consistency across extended development sessions. We also introduce an evaluation methodology with 9 metrics across comprehension and efficiency dimensions. Our framework provides agents with 8 specialized tools (file operations, search, code analysis) and evaluates them across context lengths ranging from 10K to 1M tokens, enabling precise assessment of long-context performance. Through systematic evaluation of state-of-the-art models, we reveal several key findings: (1) agents exhibit remarkable long-context robustness; (2) comprehension-efficiency trade-off exists with negative correlation, where thorough exploration increases comprehension but reduces efficiency; and (3) conversation efficiency varies dramatically across models, with strategic tool usage patterns differentiating high-performing agents. As the first long-context LLM agent benchmark for software engineering, LoCoBench-Agent establishes a rigorous foundation for measuring agent capabilities, identifying performance gaps, and advancing autonomous software development at scale.

High-dimensional decision-making tasks, such as business partner selection, involve evaluating large candidate pools with heterogeneous numerical, categorical, and textual features. While large language models (LLMs) offer strong in-context reasoning capabilities, single-agent or debate-style systems often struggle with scalability and consistency in such settings. We propose PartnerMAS, a hierarchical multi-agent framework that decomposes evaluation into three layers: a Planner Agent that designs strategies, Specialized Agents that perform role-specific assessments, and a Supervisor Agent that integrates their outputs. To support systematic evaluation, we also introduce a curated benchmark dataset of venture capital co-investments, featuring diverse firm attributes and ground-truth syndicates. Across 140 cases, PartnerMAS consistently outperforms single-agent and debate-based multi-agent baselines, achieving up to 10--15\% higher match rates. Analysis of agent reasoning shows that planners are most responsive to domain-informed prompts, specialists produce complementary feature coverage, and supervisors play an important role in aggregation. Our findings demonstrate that structured collaboration among LLM agents can generate more robust outcomes than scaling individual models, highlighting PartnerMAS as a promising framework for high-dimensional decision-making in data-rich domains.

Recent advancements in Large Language Models have sparked interest in their potential for robotic task planning. While these models demonstrate strong generative capabilities, their effectiveness in producing structured and executable plans remains uncertain. This paper presents a systematic evaluation of a broad spectrum of current state of the art language models, each directly prompted using Planning Domain Definition Language domain and problem files, and compares their planning performance with the Fast Downward planner across a variety of benchmarks. In addition to measuring success rates, we assess how faithfully the generated plans translate into sequences of actions that can actually be executed, identifying both strengths and limitations of using these models in this setting. Our findings show that while the models perform well on simpler planning tasks, they continue to struggle with more complex scenarios that require precise resource management, consistent state tracking, and strict constraint compliance. These results underscore fundamental challenges in applying language models to robotic planning in real world environments. By outlining the gaps that emerge during execution, we aim to guide future research toward combined approaches that integrate language models with classical planners in order to enhance the reliability and scalability of planning in autonomous robotics.

Post-training is essential for refining the domain-specific capabilities of large language models (LLMs), yet existing reward optimization and distribution matching methods tightly couple policy exploration with distribution alignment. This coupling forces expensive exploration directly on the policy model and severely hinders the asynchronous generation, reuse, and cross-model transfer of optimization signals. In this paper, we propose Proxy-guided Update Signal Transfer (PUST), a novel post-training framework that fundamentally decouples update-signal exploration from distribution alignment. Instead of utilizing the primary model for costly exploration, PUST employs a lightweight proxy model as an efficient testbed to discover high-reward behaviors. We extract the relative improvement signal between the proxy's initial and optimized states, transferring this directional update to the primary model to guide its policy alignment. This decoupled pipeline, comprising proxy exploration, update-signal extraction, and signal transfer, significantly reduces computational overhead and enables optimization signals to be asynchronously generated, cached, and reused. Crucially, by transferring relative improvements rather than absolute policy distributions, PUST naturally supports weak-to-strong improvement and seamless cross-model transfer. Systematic evaluations on Qwen3-family models across math and code domains demonstrate that update signals extracted from substantially weaker proxies can robustly and adjustably enhance stronger primary models. Ultimately, PUST transforms post-training from a monolithic online optimization process into a highly modular, reusable, and cost-efficient paradigm.

Recent advances in large language models (LLMs) have opened new opportunities for academic literature retrieval. However, existing systems often rely on rigid pipelines and exhibit limited reasoning capabilities. We introduce SPAR, a multi-agent framework that incorporates RefChain-based query decomposition and query evolution to enable more flexible and effective search. To facilitate systematic evaluation, we also construct SPARBench, a challenging benchmark with expert-annotated relevance labels. Experimental results demonstrate that SPAR substantially outperforms strong baselines, achieving up to +56% F1 on AutoScholar and +23% F1 on SPARBench over the best-performing baseline. Together, SPAR and SPARBench provide a scalable, interpretable, and high-performing foundation for advancing research in scholarly retrieval. Code and data will be available at: https://github.com/xiaofengShi/SPAR

We introduce LongDA, a data analysis benchmark for evaluating LLM-based agents under documentation-intensive analytical workflows. In contrast to existing benchmarks that assume well-specified schemas and inputs, LongDA targets real-world settings in which navigating long documentation and complex data is the primary bottleneck. To this end, we manually curate raw data files, long and heterogeneous documentation, and expert-written publications from 17 publicly available U.S. national surveys, from which we extract 505 analytical queries grounded in real analytical practice. Solving these queries requires agents to first retrieve and integrate key information from multiple unstructured documents, before performing multi-step computations and writing executable code, which remains challenging for existing data analysis agents. To support the systematic evaluation under this setting, we develop LongTA, a tool-augmented agent framework that enables document access, retrieval, and code execution, and evaluate a range of proprietary and open-source models. Our experiments reveal substantial performance gaps even among state-of-the-art models, highlighting the challenges researchers should consider before applying LLM agents for decision support in real-world, high-stakes analytical settings.

As large language models (LLMs) grow in scale and specialization, routing--selecting the best model for a given input--has become essential for efficient and effective deployment. While recent methods rely on complex learned routing strategies, their dependence on disparate training data and evaluation setups makes comparison and generalization difficult. In this work, we revisit LLM routing through the lens of simplicity. We show that a well-tuned k-Nearest Neighbors (kNN) approach not only matches but often outperforms state-of-the-art learned routers across diverse tasks. To support systematic evaluation, we introduce a suite of standardized routing benchmarks spanning instruction-following, question-answering, and reasoning tasks, as well as the first multi-modal routing dataset involving visual inputs. Our findings reveal that the locality properties of model performance in embedding space enable simple non-parametric methods to achieve strong routing decisions with lower sample complexity than parametric approaches. This challenges the prevailing trend toward sophisticated architectures and highlights the importance of thoroughly evaluating simple baselines before investing in complex solutions. To support reproducibility and further exploration, we will release all benchmarks and code upon publication.

Large Language Models (LLMs) have become integral to numerous domains, significantly advancing applications in data management, mining, and analysis. Their profound capabilities in processing and interpreting complex language data, however, bring to light pressing concerns regarding data privacy, especially the risk of unintentional training data leakage. Despite the critical nature of this issue, there has been no existing literature to offer a comprehensive assessment of data privacy risks in LLMs. Addressing this gap, our paper introduces LLM-PBE, a toolkit crafted specifically for the systematic evaluation of data privacy risks in LLMs. LLM-PBE is designed to analyze privacy across the entire lifecycle of LLMs, incorporating diverse attack and defense strategies, and handling various data types and metrics. Through detailed experimentation with multiple LLMs, LLM-PBE facilitates an in-depth exploration of data privacy concerns, shedding light on influential factors such as model size, data characteristics, and evolving temporal dimensions. This study not only enriches the understanding of privacy issues in LLMs but also serves as a vital resource for future research in the field. Aimed at enhancing the breadth of knowledge in this area, the findings, resources, and our full technical report are made available at https://llm-pbe.github.io/, providing an open platform for academic and practical advancements in LLM privacy assessment.

Prompt leakage poses a compelling security and privacy threat in LLM applications. Leakage of system prompts may compromise intellectual property, and act as adversarial reconnaissance for an attacker. A systematic evaluation of prompt leakage threats and mitigation strategies is lacking, especially for multi-turn LLM interactions. In this paper, we systematically investigate LLM vulnerabilities against prompt leakage for 10 closed- and open-source LLMs, across four domains. We design a unique threat model which leverages the LLM sycophancy effect and elevates the average attack success rate (ASR) from 17.7% to 86.2% in a multi-turn setting. Our standardized setup further allows dissecting leakage of specific prompt contents such as task instructions and knowledge documents. We measure the mitigation effect of 7 black-box defense strategies, along with finetuning an open-source model to defend against leakage attempts. We present different combination of defenses against our threat model, including a cost analysis. Our study highlights key takeaways for building secure LLM applications and provides directions for research in multi-turn LLM interactions

Large language models (LLMs) have achieved remarkable success in NLP and multimodal tasks. Despite these successes, their development faces two main challenges: (i) high computational cost; and (ii) difficulty in conducting fair and objective evaluations. LLMs are prohibitively expensive, making it feasible for only a few major players to undertake their training, thereby constraining both research and application opportunities. This underscores the importance of cost-effective LLM training. In this paper, we utilize a growth strategy to significantly reduce LLM training cost. We demonstrate that an LLM with 101B parameters and 0.31TB tokens can be trained on a 100K budget. We also adopt a systematic evaluation paradigm for the IQ evaluation of LLMs, in complement to existing evaluations that focus more on knowledge-oriented abilities. We introduce our benchmark including evaluations on important aspects of intelligence including symbolic mapping, itrule understanding, pattern mining, and anti-interference. Such evaluations minimize the potential impact of memorization. Experimental results show that our model FLM-101B, trained with a budget of 100K, achieves comparable performance to powerful and well-known models, eg GPT-3 and GLM-130B, especially in the IQ benchmark evaluations with contexts unseen in training data. The checkpoint of FLM-101B will be open-sourced at https://huggingface.co/CofeAI/FLM-101B.

Evaluating large language models (LLMs) for medical applications remains challenging due to benchmark saturation, limited data accessibility, and insufficient coverage of relevant tasks. Existing suites have either saturated, heavily depend on restricted datasets, or lack comprehensive model coverage. We introduce Medmarks, a fully open-source evaluation suite with 30 benchmarks spanning question answering, information extraction, medical calculations, and open-ended clinical reasoning. We perform a systematic evaluation of 61 models across 71 configurations using verifiable metrics and LLM-as-a-Judge. Our results show that frontier reasoning models (Gemini 3 Pro Preview, GPT-5.1, & GPT-5.2) achieve the highest performance across both benchmarks, most frontier proprietary models are significantly more token efficient than open-weight alternatives, medically fine-tuned models outperform their generalist counterparts, and that models are susceptible to answer-order bias (particularly smaller models and Grok 4). A subset of our evals (Medmarks-T) can be directly used as reinforcement learning environments to post-train LLMs for medical reasoning. Code is available at https://github.com/MedARC-AI/Medmarks

Multi-agent systems powered by large language models (LLMs) are transforming enterprise automation, yet systematic evaluation methodologies for assessing tool-use reliability remain underdeveloped. We introduce a comprehensive diagnostic framework that leverages big data analytics to evaluate procedural reliability in intelligent agent systems, addressing critical needs for SME-centric deployment in privacy-sensitive environments. Our approach features a 12-category error taxonomy capturing failure modes across tool initialization, parameter handling, execution, and result interpretation. Through systematic evaluation of 1,980 deterministic test instances spanning both open-weight models (Qwen2.5 series, Functionary) and proprietary alternatives (GPT-4, Claude 3.5/3.7) across diverse edge hardware configurations, we identify actionable reliability thresholds for production deployment. Our analysis reveals that procedural reliability, particularly tool initialization failures, constitutes the primary bottleneck for smaller models, while qwen2.5:32b achieves flawless performance matching GPT-4.1. The framework demonstrates that mid-sized models (qwen2.5:14b) offer practical accuracy-efficiency trade-offs on commodity hardware (96.6\% success rate, 7.3 s latency), enabling cost-effective intelligent agent deployment for resource-constrained organizations. This work establishes foundational infrastructure for systematic reliability evaluation of tool-augmented multi-agent AI systems.

Recent advances in large language models (LLMs) for mathematical reasoning have largely focused on tasks with easily verifiable final answers; however, generating and verifying natural language math proofs remains an open challenge. We identify the absence of a reliable, fine-grained evaluator for LLM-generated math proofs as a critical gap. To address this, we propose a systematic methodology for developing and validating evaluators that assign fine-grained scores on a 0-7 scale to model-generated math proofs. To enable this study, we introduce ProofBench, the first expert-annotated dataset of fine-grained proof ratings, spanning 145 problems from six major math competitions (USAMO, IMO, Putnam, etc) and 435 LLM-generated solutions from Gemini-2.5-pro, o3, and DeepSeek-R1. %with expert gradings. Using ProofBench as a testbed, we systematically explore the evaluator design space across key axes: the backbone model, input context, instructions and evaluation workflow. Our analysis delivers ProofGrader, an evaluator that combines a strong reasoning backbone LM, rich context from reference solutions and marking schemes, and a simple ensembling method; it achieves a low Mean Absolute Error (MAE) of 0.926 against expert scores, significantly outperforming naive baselines. Finally, we demonstrate its practical utility in a best-of-n selection task: at n=16, ProofGrader achieves an average score of 4.14 (out of 7), closing 78% of the gap between a naive binary evaluator (2.48) and the human oracle (4.62), highlighting its potential to advance downstream proof generation.

Benchmarks are necessary for healthcare evaluation, but are not sufficient for predicting deployment performance. Our position is that the evaluation--deployment gap arises not because of poorly designed benchmarks, but from implicit assumptions about how users interact with models that cannot be surfaced from benchmarks alone. To make this precise, we propose a classification of assumptions into two categories: task, which can be tested from conversation data alone, and outcome, which requires outcome data and behavioral studies for testing. Critically, outcome assumptions depend on human behavior, something that even well-designed benchmarks cannot directly observe. To demonstrate the operationality of this framework, we retrospectively analyze a healthcare RCT as a case study and find that the gap naturally separates into task and outcome gaps of roughly equal size. To address this, we make two contributions: first, we propose BenchmarkCards, an artifact that documents assumptions, and second, we propose staged evaluation, a procedure that systematically tests assumptions and evaluates performance.

As large language models (LLMs) become increasingly deployed in safety-critical applications, the lack of systematic methods to assess their vulnerability to jailbreak attacks presents a critical security gap. We introduce the jailbreak oracle problem: given a model, prompt, and decoding strategy, determine whether a jailbreak response can be generated with likelihood exceeding a specified threshold. This formalization enables a principled study of jailbreak vulnerabilities. Answering the jailbreak oracle problem poses significant computational challenges -- the search space grows exponentially with the length of the response tokens. We present Boa, the first efficient algorithm for solving the jailbreak oracle problem. Boa employs a three-phase search strategy: (1) constructing block lists to identify refusal patterns, (2) breadth-first sampling to identify easily accessible jailbreaks, and (3) depth-first priority search guided by fine-grained safety scores to systematically explore promising low-probability paths. Boa enables rigorous security assessments including systematic defense evaluation, standardized comparison of red team attacks, and model certification under extreme adversarial conditions.

Large Language Models (LLMs) have recently been leveraged for asset pricing tasks and stock trading applications, enabling AI agents to generate investment decisions from unstructured financial data. However, most evaluations of LLM timing-based investing strategies are conducted on narrow timeframes and limited stock universes, overstating effectiveness due to survivorship and data-snooping biases. We critically assess their generalizability and robustness by proposing FINSABER, a backtesting framework evaluating timing-based strategies across longer periods and a larger universe of symbols. Systematic backtests over two decades and 100+ symbols reveal that previously reported LLM advantages deteriorate significantly under broader cross-section and over a longer-term evaluation. Our market regime analysis further demonstrates that LLM strategies are overly conservative in bull markets, underperforming passive benchmarks, and overly aggressive in bear markets, incurring heavy losses. These findings highlight the need to develop LLM strategies that are able to prioritise trend detection and regime-aware risk controls over mere scaling of framework complexity.

The adoption of Large Language Models (LLMs) as automated evaluators (LLM-as-a-judge) has revealed critical inconsistencies in current evaluation frameworks. We identify two fundamental types of inconsistencies: (1) Score-Comparison Inconsistency, where lower-rated responses outperform higher-scored ones in pairwise comparisons, and (2) Pairwise Transitivity Inconsistency, manifested through circular preference chains (A>B>C>A) and equivalence contradictions (A=B=C\neq A). We argue that these issues come from information loss in discrete rating systems and ambiguous tie judgments during pairwise evaluation. We propose TrustJudge, a probabilistic framework that addresses these limitations through two key innovations: 1) distribution-sensitive scoring that computes continuous expectations from discrete rating probabilities, preserving information entropy for more precise scoring, and 2) likelihood-aware aggregation that resolves transitivity violations using bidirectional preference probabilities or perplexity. We also formalize the theoretical limitations of current LLM-as-a-judge frameworks and demonstrate how TrustJudge's components overcome them. When evaluated with Llama-3.1-70B-Instruct as judge using our dataset, TrustJudge reduces Score-Comparison inconsistency by 8.43% (from 23.32% to 14.89%) and Pairwise Transitivity inconsistency by 10.82% (from 15.22% to 4.40%), while maintaining higher evaluation accuracy. Our work provides the first systematic analysis of evaluation framework inconsistencies in LLM-as-a-judge paradigms, offering both theoretical insights and practical solutions for reliable automated assessment. The framework demonstrates consistent improvements across various model architectures and scales, enabling more trustworthy LLM evaluation without requiring additional training or human annotations. The codes can be found at https://github.com/TrustJudge/TrustJudge.

Evaluation and alignment pipelines for large language models increasingly rely on LLM-based judges, whose behavior is guided by natural-language rubrics and validated on benchmarks. We identify a previously under-recognized vulnerability in this workflow, which we term Rubric-Induced Preference Drift (RIPD). Even when rubric edits pass benchmark validation, they can still produce systematic and directional shifts in a judge's preferences on target domains. Because rubrics serve as a high-level decision interface, such drift can emerge from seemingly natural, criterion-preserving edits and remain difficult to detect through aggregate benchmark metrics or limited spot-checking. We further show this vulnerability can be exploited through rubric-based preference attacks, in which benchmark-compliant rubric edits steer judgments away from a fixed human or trusted reference on target domains, systematically inducing RIPD and reducing target-domain accuracy up to 9.5% (helpfulness) and 27.9% (harmlessness). When these judgments are used to generate preference labels for downstream post-training, the induced bias propagates through alignment pipelines and becomes internalized in trained policies. This leads to persistent and systematic drift in model behavior. Overall, our findings highlight evaluation rubrics as a sensitive and manipulable control interface, revealing a system-level alignment risk that extends beyond evaluator reliability alone. The code is available at: https://github.com/ZDCSlab/Rubrics-as-an-Attack-Surface. Warning: Certain sections may contain potentially harmful content that may not be appropriate for all readers.

Transformers have revolutionized the machine learning landscape, gradually making their way into everyday tasks and equipping our computers with "sparks of intelligence". However, their runtime requirements have prevented them from being broadly deployed on mobile. As personal devices become increasingly powerful and prompt privacy becomes an ever more pressing issue, we explore the current state of mobile execution of Large Language Models (LLMs). To achieve this, we have created our own automation infrastructure, MELT, which supports the headless execution and benchmarking of LLMs on device, supporting different models, devices and frameworks, including Android, iOS and Nvidia Jetson devices. We evaluate popular instruction fine-tuned LLMs and leverage different frameworks to measure their end-to-end and granular performance, tracing their memory and energy requirements along the way. Our analysis is the first systematic study of on-device LLM execution, quantifying performance, energy efficiency and accuracy across various state-of-the-art models and showcases the state of on-device intelligence in the era of hyperscale models. Results highlight the performance heterogeneity across targets and corroborates that LLM inference is largely memory-bound. Quantization drastically reduces memory requirements and renders execution viable, but at a non-negligible accuracy cost. Drawing from its energy footprint and thermal behavior, the continuous execution of LLMs remains elusive, as both factors negatively affect user experience. Last, our experience shows that the ecosystem is still in its infancy, and algorithmic as well as hardware breakthroughs can significantly shift the execution cost. We expect NPU acceleration, and framework-hardware co-design to be the biggest bet towards efficient standalone execution, with the alternative of offloading tailored towards edge deployments.

We present a systematic, empirical evaluation of five local large language model (LLM) runtimes on Apple Silicon: MLX, MLC-LLM, llama.cpp, Ollama, and PyTorch MPS. Experiments were conducted on a Mac Studio equipped with an M2 Ultra processor and 192 GB of unified memory. Using the Qwen-2.5 model family across prompts ranging from a few hundred to 100,000 tokens, we measure time-to-first-token (TTFT), steady-state throughput, latency percentiles, long-context behavior (key-value and prompt caching), quantization support, streaming performance, batching and concurrency behavior, and deployment complexity. Under our settings, MLX achieves the highest sustained generation throughput, while MLC-LLM delivers consistently lower TTFT for moderate prompt sizes and offers stronger out-of-the-box inference features. llama.cpp is highly efficient for lightweight single-stream use, Ollama emphasizes developer ergonomics but lags in throughput and TTFT, and PyTorch MPS remains limited by memory constraints on large models and long contexts. All frameworks execute fully on-device with no telemetry, ensuring strong privacy guarantees. We release scripts, logs, and plots to reproduce all results. Our analysis clarifies the design trade-offs in Apple-centric LLM deployments and provides evidence-based recommendations for interactive and long-context processing. Although Apple Silicon inference frameworks still trail NVIDIA GPU-based systems such as vLLM in absolute performance, they are rapidly maturing into viable, production-grade solutions for private, on-device LLM inference.

Large language models (LLMs) have emerged as a new paradigm for Text-to-SQL task. However, the absence of a systematical benchmark inhibits the development of designing effective, efficient and economic LLM-based Text-to-SQL solutions. To address this challenge, in this paper, we first conduct a systematical and extensive comparison over existing prompt engineering methods, including question representation, example selection and example organization, and with these experimental results, we elaborate their pros and cons. Based on these findings, we propose a new integrated solution, named DAIL-SQL, which refreshes the Spider leaderboard with 86.6% execution accuracy and sets a new bar. To explore the potential of open-source LLM, we investigate them in various scenarios, and further enhance their performance with supervised fine-tuning. Our explorations highlight open-source LLMs' potential in Text-to-SQL, as well as the advantages and disadvantages of the supervised fine-tuning. Additionally, towards an efficient and economic LLM-based Text-to-SQL solution, we emphasize the token efficiency in prompt engineering and compare the prior studies under this metric. We hope that our work provides a deeper understanding of Text-to-SQL with LLMs, and inspires further investigations and broad applications.

Large language models are emerging as powerful tools for scientific law discovery, a foundational challenge in AI-driven science. However, existing benchmarks for this task suffer from a fundamental methodological trilemma, forcing a trade-off between scientific relevance, scalability, and resistance to memorization. Furthermore, they oversimplify discovery as static function fitting, failing to capture the authentic scientific process of uncovering embedded laws through the interactive exploration of complex model systems. To address these critical gaps, we introduce NewtonBench, a benchmark comprising 324 scientific law discovery tasks across 12 physics domains. Our design mitigates the evaluation trilemma by using metaphysical shifts - systematic alterations of canonical laws - to generate a vast suite of problems that are scalable, scientifically relevant, and memorization-resistant. Moreover, we elevate the evaluation from static function fitting to interactive model discovery, requiring agents to experimentally probe simulated complex systems to uncover hidden principles. Our extensive experiment reveals a clear but fragile capability for discovery in frontier LLMs: this ability degrades precipitously with increasing system complexity and exhibits extreme sensitivity to observational noise. Notably, we uncover a paradoxical effect of tool assistance: providing a code interpreter can hinder more capable models by inducing a premature shift from exploration to exploitation, causing them to satisfice on suboptimal solutions. These results demonstrate that robust, generalizable discovery in complex, interactive environments remains the core challenge. By providing a scalable, robust, and scientifically authentic testbed, NewtonBench offers a crucial tool for measuring true progress and guiding the development of next-generation AI agents capable of genuine scientific discovery.

Many companies use large language models (LLMs) offered as a service, like OpenAI's GPT-4, to create AI-enabled product experiences. Along with the benefits of ease-of-use and shortened time-to-solution, this reliance on proprietary services has downsides in model control, performance reliability, uptime predictability, and cost. At the same time, a flurry of open-source small language models (SLMs) has been made available for commercial use. However, their readiness to replace existing capabilities remains unclear, and a systematic approach to holistically evaluate these SLMs is not readily available. This paper presents a systematic evaluation methodology and a characterization of modern open-source SLMs and their trade-offs when replacing proprietary LLMs for a real-world product feature. We have designed SLaM, an open-source automated analysis tool that enables the quantitative and qualitative testing of product features utilizing arbitrary SLMs. Using SLaM, we examine the quality and performance characteristics of modern SLMs relative to an existing customer-facing implementation using the OpenAI GPT-4 API. Across 9 SLMs and their 29 variants, we observe that SLMs provide competitive results, significant performance consistency improvements, and a cost reduction of 5x~29x when compared to GPT-4.

Large language models (LLMs) can generate BPMN process models from natural-language descriptions, yet supervised fine-tuning (SFT) limits their output quality to the patterns present in the training data. Reinforcement learning (RL) can optimize beyond this ceiling using external quality measures, but how the reward function should be designed when quality is multi-dimensional remains unexplored. We present a systematic investigation of reward function design for RL-based process model generation, training two LLM families (Llama~3.1 8B, Qwen~2.5 14B) under 48 configurations using Group Sequence Policy Optimization with rewards derived from an automated evaluation framework comprising 38 metrics across syntactic, pragmatic, and semantic quality. Three findings emerge. First, RL significantly improves pragmatic and syntactic quality while preserving semantic fidelity, reducing output variability by more than sixfold. Second, equal reward weighting consistently outperforms targeted weighting: emphasizing a specific dimension fails to improve it and can collapse the model into a low-quality mode. Third, design choices interact with model architecture in non-trivial ways: the invalidity penalty is essential for one model but irrelevant for the other, and SFT initialization is indispensable for one architecture but counterproductive for another. These results demonstrate that reward composition is a primary determinant of optimization outcomes, with effects as large as the decision to apply RL itself. The findings generalize to any structured generation task where quality is assessed along multiple automated dimensions. We release our implementation and experimental code at https://github.com/chlauer99/RL_for_process_modeling.

Large language model agents equipped with persistent memory are vulnerable to memory poisoning attacks, where adversaries inject malicious instructions through query only interactions that corrupt the agents long term memory and influence future responses. Recent work demonstrated that the MINJA (Memory Injection Attack) achieves over 95 % injection success rate and 70 % attack success rate under idealized conditions. However, the robustness of these attacks in realistic deployments and effective defensive mechanisms remain understudied. This work addresses these gaps through systematic empirical evaluation of memory poisoning attacks and defenses in Electronic Health Record (EHR) agents. We investigate attack robustness by varying three critical dimensions: initial memory state, number of indication prompts, and retrieval parameters. Our experiments on GPT-4o-mini, Gemini-2.0-Flash and Llama-3.1-8B-Instruct models using MIMIC-III clinical data reveal that realistic conditions with pre-existing legitimate memories dramatically reduce attack effectiveness. We then propose and evaluate two novel defense mechanisms: (1) Input/Output Moderation using composite trust scoring across multiple orthogonal signals, and (2) Memory Sanitization with trust-aware retrieval employing temporal decay and pattern-based filtering. Our defense evaluation reveals that effective memory sanitization requires careful trust threshold calibration to prevent both overly conservative rejection (blocking all entries) and insufficient filtering (missing subtle attacks), establishing important baselines for future adaptive defense mechanisms. These findings provide crucial insights for securing memory-augmented LLM agents in production environments.

Adversarial board games, as a paradigmatic domain of strategic reasoning and intelligence, have long served as both a popular competitive activity and a benchmark for evaluating artificial intelligence (AI) systems. Building on this foundation, we propose an adversarial benchmarking framework to assess the comprehensive performance of Large Language Models (LLMs) through board games competition, compensating the limitation of data dependency of the mainstream Question-and-Answer (Q&A) based benchmark method. We introduce Qi Town, a specialized evaluation platform that supports 5 widely played games and involves 20 LLM-driven players. The platform employs both the Elo rating system and a novel Performance Loop Graph (PLG) to quantitatively evaluate the technical capabilities of LLMs, while also capturing Positive Sentiment Score (PSS) throughout gameplay to assess mental fitness. The evaluation is structured as a round-robin tournament, enabling systematic comparison across players. Experimental results indicate that, despite technical differences, most LLMs remain optimistic about winning and losing, demonstrating greater adaptability to high-stress adversarial environments than humans. On the other hand, the complex relationship between cyclic wins and losses in PLGs exposes the instability of LLMs' skill play during games, warranting further explanation and exploration.

Rubric-based scoring has become a widely used paradigm in model evaluation, typically with LLM-as-a-Judge (LaaJ) for rubric scoring. However, the reliability of LaaJ for rubric scoring remains underexplored. This concern is especially pronounced in agentic scenarios, where long, complex outputs further challenge reliable scoring. To address this, we conduct a systematic meta-evaluation of LaaJ reliability for rubric verification. We introduce RuVerBench, the first benchmark for assessing LaaJ reliability in rubric verification for agentic scenarios. RuVerBench covers two prevalent agentic domains, deep research and agentic coding, with 2,458 instances, each containing a model-generated output, a rubric, and a human-annotated label indicating whether the output satisfies the rubric. Using RuVerBench, we evaluate numerous frontier LLMs and find that even the most advanced models achieve strong performance but still exhibit substantial noise. We further analyze the impact of key LaaJ strategies, including prompt design, batching, and majority voting, on rubric verification. We find that weaker models are more sensitive to prompt variations, batched verification presents a trade-off between accuracy and efficiency, and majority voting yields effective but diminishing returns. We have released our dataset and code to facilitate future research: https://github.com/THU-KEG/RuVerBench.

We introduce Earth Virtual Expert (EVE), the first open-source, end-to-end initiative for developing and deploying domain-specialized LLMs for Earth Intelligence. At its core is EVE-Instruct, a domain-adapted 24B model built on Mistral Small 3.2 and optimized for reasoning and question answering. On newly constructed Earth Observation and Earth Sciences benchmarks, it outperforms comparable models while preserving general capabilities. We release curated training corpora and the first systematic domain-specific evaluation benchmarks, covering MCQA, open-ended QA, and factuality. EVE further integrates RAG and a hallucination-detection pipeline into a production system deployed via API and GUI, supporting 350 pilot users so far. All models, datasets, and code are ready to be released under open licenses as contributions to our field at huggingface.co/eve-esa and github.com/eve-esa.

The rapid adoption of LLM-based agentic systems has produced a rich ecosystem of frameworks (smolagents, LangGraph, AutoGen, CAMEL, LlamaIndex, i.a.). Yet existing benchmarks are model-centric: they fix the agentic setup and do not compare other system components. We argue that implementation decisions substantially impact performance, including choices such as topology, orchestration logic, and error handling. MASEval addresses this evaluation gap with a framework-agnostic library that treats the entire system as the unit of analysis. Through a systematic system-level comparison across 3 benchmarks, 3 models, and 3 frameworks, we find that framework choice matters as much as model choice. MASEval allows researchers to explore all components of agentic systems, opening new avenues for principled system design, and practitioners to identify the best implementation for their use case. MASEval is available under the MIT licence https://github.com/parameterlab/MASEval.

While large language models (LLMs) achieve near-perfect scores on medical licensing exams, these evaluations inadequately reflect the complexity and diversity of real-world clinical practice. We introduce MedHELM, an extensible evaluation framework for assessing LLM performance for medical tasks with three key contributions. First, a clinician-validated taxonomy spanning 5 categories, 22 subcategories, and 121 tasks developed with 29 clinicians. Second, a comprehensive benchmark suite comprising 35 benchmarks (17 existing, 18 newly formulated) providing complete coverage of all categories and subcategories in the taxonomy. Third, a systematic comparison of LLMs with improved evaluation methods (using an LLM-jury) and a cost-performance analysis. Evaluation of 9 frontier LLMs, using the 35 benchmarks, revealed significant performance variation. Advanced reasoning models (DeepSeek R1: 66% win-rate; o3-mini: 64% win-rate) demonstrated superior performance, though Claude 3.5 Sonnet achieved comparable results at 40% lower estimated computational cost. On a normalized accuracy scale (0-1), most models performed strongly in Clinical Note Generation (0.73-0.85) and Patient Communication & Education (0.78-0.83), moderately in Medical Research Assistance (0.65-0.75), and generally lower in Clinical Decision Support (0.56-0.72) and Administration & Workflow (0.53-0.63). Our LLM-jury evaluation method achieved good agreement with clinician ratings (ICC = 0.47), surpassing both average clinician-clinician agreement (ICC = 0.43) and automated baselines including ROUGE-L (0.36) and BERTScore-F1 (0.44). Claude 3.5 Sonnet achieved comparable performance to top models at lower estimated cost. These findings highlight the importance of real-world, task-specific evaluation for medical use of LLMs and provides an open source framework to enable this.

The emergence of large language models (LLMs) has significantly influenced numerous fields, including healthcare, by enhancing the capabilities of automated systems to process and generate human-like text. However, despite their advancements, the reliability and accuracy of LLMs in medical contexts remain critical concerns. Current evaluation methods often lack robustness and fail to provide a comprehensive assessment of LLM performance, leading to potential risks in clinical settings. In this work, we propose Med-CoDE, a specifically designed evaluation framework for medical LLMs to address these challenges. The framework leverages a critique-based approach to quantitatively measure the degree of disagreement between model-generated responses and established medical ground truths. This framework captures both accuracy and reliability in medical settings. The proposed evaluation framework aims to fill the existing gap in LLM assessment by offering a systematic method to evaluate the quality and trustworthiness of medical LLMs. Through extensive experiments and case studies, we illustrate the practicality of our framework in providing a comprehensive and reliable evaluation of medical LLMs.

As large language models (LLMs) are increasingly deployed for text generation, watermarking has become essential for authorship attribution, intellectual property protection, and misuse detection. While existing watermarking methods perform well in high-resource languages, their robustness in low-resource languages remains underexplored. This work presents the first systematic evaluation of state-of-the-art text watermarking methods: KGW, Exponential Sampling (EXP), and Waterfall, for Bangla LLM text generation under cross-lingual round-trip translation (RTT) attacks. Under benign conditions, KGW and EXP achieve high detection accuracy (>88%) with negligible perplexity and ROUGE degradation. However, RTT causes detection accuracy to collapse below RTT causes detection accuracy to collapse to 9-13%, indicating a fundamental failure of token-level watermarking. To address this, we propose a layered watermarking strategy that combines embedding-time and post-generation watermarks. Experimental results show that layered watermarking improves post-RTT detection accuracy by 25-35%, achieving 40-50% accuracy, representing a 3times to 4times relative improvement over single-layer methods, at the cost of controlled semantic degradation. Our findings quantify the robustness-quality trade-off in multilingual watermarking and establish layered watermarking as a practical, training-free solution for low-resource languages such as Bangla. Our code and data will be made public.

We propose LogicVista, an evaluation benchmark that assesses the integrated logical reasoning capabilities of multimodal large language models (MLLMs) in Visual contexts. Recent advancements in MLLMs have demonstrated various fascinating abilities, from crafting poetry based on an image to performing mathematical reasoning. However, there is still a lack of systematic evaluation of MLLMs' proficiency in logical reasoning tasks, which are essential for activities like navigation and puzzle-solving. Thus we evaluate general logical cognition abilities across 5 logical reasoning tasks encompassing 9 different capabilities, using a sample of 448 multiple-choice questions. Each question is annotated with the correct answer and the human-written reasoning behind the selection, enabling both open-ended and multiple-choice evaluation. A total of 8 MLLMs are comprehensively evaluated using LogicVista. Code and Data Available at https://github.com/Yijia-Xiao/LogicVista.

Summarizing book-length documents (>100K tokens) that exceed the context window size of large language models (LLMs) requires first breaking the input document into smaller chunks and then prompting an LLM to merge, update, and compress chunk-level summaries. Despite the complexity and importance of this task, it has yet to be meaningfully studied due to the challenges of evaluation: existing book-length summarization datasets (e.g., BookSum) are in the pretraining data of most public LLMs, and existing evaluation methods struggle to capture errors made by modern LLM summarizers. In this paper, we present the first study of the coherence of LLM-based book-length summarizers implemented via two prompting workflows: (1) hierarchically merging chunk-level summaries, and (2) incrementally updating a running summary. We obtain 1193 fine-grained human annotations on GPT-4 generated summaries of 100 recently-published books and identify eight common types of coherence errors made by LLMs. Because human evaluation is expensive and time-consuming, we develop an automatic metric, BooookScore, that measures the proportion of sentences in a summary that do not contain any of the identified error types. BooookScore has high agreement with human annotations and allows us to systematically evaluate the impact of many other critical parameters (e.g., chunk size, base LLM) while saving $15K USD and 500 hours in human evaluation costs. We find that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher BooookScore than those generated by open-source models. While LLaMA 2 falls behind other models, Mixtral achieves performance on par with GPT-3.5-Turbo. Incremental updating yields lower BooookScore but higher level of detail than hierarchical merging, a trade-off sometimes preferred by annotators.

Evaluating the quality and variability of text generated by Large Language Models (LLMs) poses a significant, yet unresolved research challenge. Traditional evaluation methods, such as ROUGE and BERTScore, which measure token similarity, often fail to capture the holistic semantic equivalence. This results in a low correlation with human judgments and intuition, which is especially problematic in high-stakes applications like healthcare and finance where reliability, safety, and robust decision-making are highly critical. This work proposes DCR, an automated framework for evaluating and improving the consistency of LLM-generated texts using a divide-conquer-reasoning approach. Unlike existing LLM-based evaluators that operate at the paragraph level, our method employs a divide-and-conquer evaluator (DCE) that breaks down the paragraph-to-paragraph comparison between two generated responses into individual sentence-to-paragraph comparisons, each evaluated based on predefined criteria. To facilitate this approach, we introduce an automatic metric converter (AMC) that translates the output from DCE into an interpretable numeric score. Beyond the consistency evaluation, we further present a reason-assisted improver (RAI) that leverages the analytical reasons with explanations identified by DCE to generate new responses aimed at reducing these inconsistencies. Through comprehensive and systematic empirical analysis, we show that our approach outperforms state-of-the-art methods by a large margin (e.g., +19.3% and +24.3% on the SummEval dataset) in evaluating the consistency of LLM generation across multiple benchmarks in semantic, factual, and summarization consistency tasks. Our approach also substantially reduces nearly 90% of output inconsistencies, showing promise for effective hallucination mitigation.

Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain. This gap poses real risks: under ambiguity, conflicting constraints, or tool failures, agents may execute unintended and irreversible actions. To close this gap, we present the first systematic evaluation framework for agentic abstention: the calibrated ability of tool-using LLM agents to recognize when not to act. At its core, AgentAbstain is a paired-task benchmark built on an agent-native taxonomy of 8 abstention scenarios across pre-execution reasoning and runtime discovery. It contains 263 paired tasks across 42 executable sandbox environments, where each pair consists of a should-act task and a should-abstain variant produced through a controlled perturbation to the instruction, tool, or environment state. To scale this paired design and resist data contamination, we propose AbstainGen, a fully automated pipeline that synthesizes sandbox environments and generates paired tasks end-to-end, validated by deterministic replay and semantic LLM judges; fresh task instances can be regenerated on demand, and three independent annotators rate 94-98% of sampled tasks as well-designed. Across 17 frontier LLMs in 4 agent harnesses, the best agent (Gemini 3.1 Pro) achieves only 59.5% paired accuracy (correct on both the act and abstain sides of each paired task). More importantly, abstention capability is largely independent of general task-solving capability, indicating that scaling task-solving alone will not close this gap. We further identify failure modes such as post-hoc abstention, in which agents execute irreversible actions before recognizing abstention triggers. Our code and dataset are open-sourced at agentabstain.github.io.

Large language models (LLMs) have become a standard for multilingual evaluation, yet they exhibit a severe systematic translationese bias. In this paper, translationese bias is characterized as LLMs systematically favoring machine-translated text over human-authored references, particularly in low-resource languages. We attribute this bias to spurious correlations with (i) latent manifold alignment with English and (ii) cross-lingual predictability. To mitigate this bias, we propose DIBJudge, a robust fine-tuning framework that learns a minimally sufficient, judgment-critical representation via variational information compression, while explicitly isolating spurious factors into the dedicated bias branch. Furthermore, we incorporate a cross-covariance penalty that explicitly suppresses statistical dependence between robust and bias representations, thereby encouraging effective disentanglement. Extensive evaluations on multilingual reward modeling benchmarks and a dedicated translationese bias evaluation suite demonstrate that the proposed DIBJudge consistently outperforms strong baselines and substantially mitigates translationese bias.

Meaning in human language is relational, context dependent, and emergent, arising from dynamic systems of signs rather than fixed word-concept mappings. In computational settings, this semiotic and interpretive complexity complicates the generation and evaluation of meaning. This article proposes an interdisciplinary framework for studying meaning in large language model (LLM) generated language by integrating semiotics and hermeneutics with qualitative research methods. We review prior scholarship on meaning and machines, examining how linguistic signs are transformed into vectorized representations in static and contextualized embedding models, and identify gaps between statistical approximation and human interpretive meaning. We then introduce the Inductive Conceptual Rating (ICR) metric, a qualitative evaluation approach grounded in inductive content analysis and reflexive thematic analysis, designed to assess semantic accuracy and meaning alignment in LLM-outputs beyond lexical similarity metrics. We apply ICR in an empirical comparison of LLM generated and human generated thematic summaries across five datasets (N = 50 to 800). While LLMs achieve high linguistic similarity, they underperform on semantic accuracy, particularly in capturing contextually grounded meanings. Performance improves with larger datasets but remains variable across models, potentially reflecting differences in the frequency and coherence of recurring concepts and meanings. We conclude by arguing for evaluation frameworks that leverage systematic qualitative interpretation practices when assessing meaning in LLM-generated outputs from reference texts.

Large Vision-Language Models (VLMs) excel at understanding and generating video descriptions but their high memory, computation, and deployment demands hinder practical use particularly for blind and low-vision (BLV) users who depend on detailed, context-aware descriptions. To study the effect of model size on accessibility-focused description quality, we evaluate SmolVLM2 variants with 500M and 2.2B parameters across two diverse datasets: AVCaps (outdoor), and Charades (indoor). In this work, we introduce two novel evaluation frameworks specifically designed for BLV accessibility assessment: the Multi-Context BLV Framework evaluating spatial orientation, social interaction, action events, and ambience contexts; and the Navigational Assistance Framework focusing on mobility-critical information. Additionally, we conduct a systematic evaluation of four different prompt design strategies and deploy both models on a smartphone, evaluating FP32 and INT8 precision variants to assess real-world performance constraints on resource-limited mobile devices.

Existing benchmarks have proven effective for assessing the performance of fully trained large language models. However, we find striking differences in the early training stages of small models, where benchmarks often fail to provide meaningful or discriminative signals. To explore how these differences arise, this competition tackles the challenge of designing scientific knowledge evaluation tasks specifically tailored for measuring early training progress of language models. Participants are invited to develop novel evaluation methodologies or adapt existing benchmarks to better capture performance differences among language models. To support this effort, we provide three pre-trained small models (0.5B, 1B, and 3B parameters), along with intermediate checkpoints sampled during training up to 200B tokens. All experiments and development work can be run on widely available free cloud-based GPU platforms, making participation accessible to researchers with limited computational resources. Submissions will be evaluated based on three criteria: the quality of the performance signal they produce, the consistency of model rankings at 1 trillion tokens of training, and their relevance to the scientific knowledge domain. By promoting the design of tailored evaluation strategies for early training, this competition aims to attract a broad range of participants from various disciplines, including those who may not be machine learning experts or have access to dedicated GPU resources. Ultimately, this initiative seeks to make foundational LLM research more systematic and benchmark-informed from the earliest phases of model development.

Large Language Models (LLMs) are increasingly deployed as the backend for a variety of real-world applications called LLM-Integrated Applications. Multiple recent works showed that LLM-Integrated Applications are vulnerable to prompt injection attacks, in which an attacker injects malicious instruction/data into the input of those applications such that they produce results as the attacker desires. However, existing works are limited to case studies. As a result, the literature lacks a systematic understanding of prompt injection attacks and their defenses. We aim to bridge the gap in this work. In particular, we propose a general framework to formalize prompt injection attacks. Existing attacks, which are discussed in research papers and blog posts, are special cases in our framework. Our framework enables us to design a new attack by combining existing attacks. Moreover, we also propose a framework to systematize defenses against prompt injection attacks. Using our frameworks, we conduct a systematic evaluation on prompt injection attacks and their defenses with 10 LLMs and 7 tasks. We hope our frameworks can inspire future research in this field. Our code is available at https://github.com/liu00222/Open-Prompt-Injection.

Recently, LLM agents have made rapid progress in improving their programming capabilities. However, existing benchmarks lack the ability to automatically evaluate from users' perspective, and also lack the explainability of the results of LLM agents' code generation capabilities. Thus, we introduce ProjectEval, a new benchmark for LLM agents project-level code generation's automated evaluation by simulating user interaction. ProjectEval is constructed by LLM with human reviewing. It has three different level inputs of natural languages or code skeletons. ProjectEval can evaluate the generated projects by user interaction simulation for execution, and by code similarity through existing objective indicators. Through ProjectEval, we find that systematic engineering project code, overall understanding of the project and comprehensive analysis capability are the keys for LLM agents to achieve practical projects. Our findings and benchmark provide valuable insights for developing more effective programming agents that can be deployed in future real-world production.