With the rapid development of large language models (LLMs), generative reward models (GRMs) have been widely adopted for reward modeling and evaluation. Previous studies have primarily focused on training specialized GRMs by optimizing them on preference datasets with the judgment correctness as supervision. While it's widely accepted that GRMs with stronger problem-solving capabilities typically exhibit superior judgment abilities, we first identify a significant solve-to-judge gap when examining individual queries. Specifically, the solve-to-judge gap refers to the phenomenon where GRMs struggle to make correct judgments on some queries (14%-37%), despite being fully capable of solving them. In this paper, we propose the Solve-to-Judge (S2J) approach to address this problem. Our comprehensive experiments demonstrate that S2J effectively reduces the solve-to-judge gap by 16.2%, thereby enhancing the model's judgment performance by 5.8%. Notably, S2J achieves state-of-the-art (SOT A) performance among GRMs built on the same base model while utilizing a significantly smaller training dataset. Moreover, S2J accomplishes this through self-evolution without relying on more powerful external models for distillation. As Large Language Models (LLMs) continue to evolve rapidly, a variety of evaluation paradigms have been proposed to accurately evaluate the quality of their responses. This is not only crucial for providing accurate reward signals in post-training (Ouyang et al., 2022; Bai et al., 2022; Wang et al., 2024a), but also important for automated evaluation and benchmark construction (Zheng et al., 2023; Dubois et al., 2024). Among them, Generative Reward Mmodels (GRMs) have been proposed as a solution, which treats evaluation as a capability of LLMs and leverages LLMs to evaluate other LLMs (Zheng et al., 2023; Li et al., 2025). Unlike scalar reward models, which only output a single numerical score (Liu et al., 2024a; Lambert et al., 2024), GRMs utilize the generative capabilities of LLMs to produce an interpretable analysis before rendering a verdict.

Large language model (LLM) powered code agents are rapidly transforming software engineering by automating tasks such as testing, debugging, and repairing, yet the security risks of their generated code have become a critical concern. Existing benchmarks have offered valuable insights but remain insufficient: they often overlook the genuine context in which vulnerabilities were introduced or adopt narrow evaluation protocols that fail to capture either functional correctness or newly introduced vulnerabilities. We therefore introduce SecureAgentBench, a benchmark of 105 coding tasks designed to rigorously evaluate code agents' capabilities in secure code generation. Each task includes (i) realistic task settings that require multi-file edits in large repositories, (ii) aligned contexts based on real-world open-source vulnerabilities with precisely identified introduction points, and (iii) comprehensive evaluation that combines functionality testing, vulnerability checking through proof-of-concept exploits, and detection of newly introduced vulnerabilities using static analysis. We evaluate three representative agents (SWE-agent, OpenHands, and Aider) with three state-of-the-art LLMs (Claude 3.7 Sonnet, GPT-4.1, and DeepSeek-V3.1). Results show that (i) current agents struggle to produce secure code, as even the best-performing one, SWE-agent supported by DeepSeek-V3.1, achieves merely 15.2% correct-and-secure solutions, (ii) some agents produce functionally correct code but still introduce vulnerabilities, including new ones not previously recorded, and (iii) adding explicit security instructions for agents does not significantly improve secure coding, underscoring the need for further research. These findings establish SecureAgentBench as a rigorous benchmark for secure code generation and a step toward more reliable software development with LLMs.

Robotic manipulation with Vision-Language-Action models requires efficient inference over long-horizon multi-modal context, where attention to dense visual tokens dominates computational cost. Existing methods optimize inference speed by reducing visual redundancy within VLA models, but they overlook the varying redundancy across robotic manipulation stages. We observe that the visual token redundancy is higher in coarse manipulation phase than in fine-grained operations, and is strongly correlated with the action dynamic. Motivated by this observation, we propose \textbf{A}ction-aware \textbf{D}ynamic \textbf{P}runing (\textbf{ADP}), a multi-modal pruning framework that integrates text-driven token selection with action-aware trajectory gating. Our method introduces a gating mechanism that conditions the pruning signal on recent action trajectories, using past motion windows to adaptively adjust token retention ratios in accordance with dynamics, thereby balancing computational efficiency and perceptual precision across different manipulation stages. Extensive experiments on the LIBERO suites and diverse real-world scenarios demonstrate that our method significantly reduces FLOPs and action inference latency (\textit{e.g.} $1.35 \times$ speed up on OpenVLA-OFT) while maintaining competitive success rates (\textit{e.g.} 25.8\% improvements with OpenVLA) compared to baselines, thereby providing a simple plug-in path to efficient robot policies that advances the efficiency and performance frontier of robotic manipulation. Our project website is: \href{https://vla-adp.github.io/}{ADP.com}.

Non-English dialogue datasets are scarce, and models are often trained or evaluated on translations of English-language dialogues, an approach which can introduce artifacts that reduce their naturalness and cultural appropriateness. This work proposes Dialogue Act Script (DAS), a structured framework for encoding, localizing, and generating multilingual dialogues from abstract intent representations. Rather than translating dialogue utterances directly, DAS enables the generation of new dialogues in the target language that are culturally and contextually appropriate. By using structured dialogue act representations, DAS supports flexible localization across languages, mitigating translationese and enabling more fluent, naturalistic conversations. Human evaluations across Italian, German, and Chinese show that DAS-generated dialogues consistently outperform those produced by both machine and human translators on measures of cultural relevance, coherence, and situational appropriateness.

Activation steering is a promising technique for controlling LLM behavior by adding semantically meaningful vectors directly into a model's hidden states during inference. It is often framed as a precise, interpretable, and potentially safer alternative to fine-tuning. We demonstrate the opposite: steering systematically breaks model alignment safeguards, making it comply with harmful requests. Through extensive experiments on different model families, we show that even steering in a random direction can increase the probability of harmful compliance from 0% to 2-27%. Alarmingly, steering benign features from a sparse autoencoder (SAE), a common source of interpretable directions, increases these rates by a further 2-4%. Finally, we show that combining 20 randomly sampled vectors that jailbreak a single prompt creates a universal attack, significantly increasing harmful compliance on unseen requests. These results challenge the paradigm of safety through interpretability, showing that precise control over model internals does not guarantee precise control over model behavior.

Prior work has shown that two NLP evaluation experiments that report results for the same quality criterion name (e.g. Fluency) do not necessarily evaluate the same aspect of quality, and the comparability implied by the name can be misleading. Not knowing when two evaluations are comparable in this sense means we currently lack the ability to draw reliable conclusions about system quality on the basis of multiple, independently conducted evaluations. This in turn hampers the ability of the field to progress scientifically as a whole, a pervasive issue in NLP since its beginning (Sparck Jones, 1981). It is hard to see how the issue of unclear comparability can be fully addressed other than by the creation of a standard set of quality criterion names and definitions that the several hundred quality criterion names actually in use in the field can be mapped to, and grounded in. Taking a strictly descriptive approach, the QCET Quality Criteria for Evaluation Taxonomy derives a standard set of quality criterion names and definitions from three surveys of evaluations reported in NLP, and structures them into a hierarchy where each parent node captures common aspects of its child nodes. We present QCET and the resources it consists of, and discuss its three main uses in (i) establishing comparability of existing evaluations, (ii) guiding the design of new evaluations, and (iii) assessing regulatory compliance.

Speech Continuation (SC) is the task of generating a coherent extension of a spoken prompt while preserving both semantic context and speaker identity. Because SC is constrained to a single audio stream, it offers a more direct setting for probing biases in speech foundation models than dialogue does. In this work we present the first systematic evaluation of bias in SC, investigating how gender and phonation type (breathy, creaky, end-creak) affect continuation behaviour. We evaluate three recent models: SpiritLM (base and expressive), VAE-GSLM, and SpeechGPT across speaker similarity, voice quality preservation, and text-based bias metrics. Results show that while both speaker similarity and coherence remain a challenge, textual evaluations reveal significant model and gender interactions: once coherence is sufficiently high (for VAE-GSLM), gender effects emerge on text-metrics such as agency and sentence polarity. In addition, continuations revert toward modal phonation more strongly for female prompts than for male ones, revealing a systematic voice-quality bias. These findings highlight SC as a controlled probe of socially relevant representational biases in speech foundation models, and suggest that it will become an increasingly informative diagnostic as continuation quality improves.

The proliferation of Large Language Models (LLMs) has led to widespread AI-Generated Text (AIGT) on social media platforms, creating unique challenges where content dynamics are driven by user engagement and evolve over time. However, existing datasets mainly depict static AIGT detection. In this work, we introduce RedNote-Vibe, the first longitudinal (5-years) dataset for social media AIGT analysis. This dataset is sourced from Xiaohongshu platform, containing user engagement metrics (e.g., likes, comments) and timestamps spanning from the pre-LLM period to July 2025, which enables research into the temporal dynamics and user interaction patterns of AIGT. Furthermore, to detect AIGT in the context of social media, we propose PsychoLinguistic AIGT Detection Framework (PLAD), an interpretable approach that leverages psycholinguistic features. Our experiments show that PLAD achieves superior detection performance and provides insights into the signatures distinguishing human and AI-generated content. More importantly, it reveals the complex relationship between these linguistic features and social media engagement. The dataset is available at https://github.com/testuser03158/RedNote-Vibe.

With the rapid advancement of large language models (LLMs), natural language processing (NLP) has achieved remarkable progress. Nonetheless, significant challenges remain in handling texts with ambiguity, polysemy, or uncertainty. We introduce the Fuzzy Reasoning Chain (FRC) framework, which integrates LLM semantic priors with continuous fuzzy membership degrees, creating an explicit interaction between probability-based reasoning and fuzzy membership reasoning. This transition allows ambiguous inputs to be gradually transformed into clear and interpretable decisions while capturing conflicting or uncertain signals that traditional probability-based methods cannot. We validate FRC on sentiment analysis tasks, where both theoretical analysis and empirical results show that it ensures stable reasoning and facilitates knowledge transfer across different model scales. These findings indicate that FRC provides a general mechanism for managing subtle and ambiguous expressions with improved interpretability and robustness.

Recent Large Reasoning Models have achieved significant improvements in complex task-solving capabilities by allocating more computation at the inference stage with a "thinking longer" paradigm. Even as the foundational reasoning capabilities of models advance rapidly, the persistent gap between a model's performance in a single attempt and its latent potential, often revealed only across multiple solution paths, starkly highlights the disparity between its realized and inherent capabilities. To address this, we present A2R, an Asymmetric Two-Stage Reasoning framework designed to explicitly bridge the gap between a model's potential and its actual performance. In this framework, an "explorer" model first generates potential solutions in parallel through repeated sampling. Subsequently,a "synthesizer" model integrates these references for a more refined, second stage of reasoning. This two-stage process allows computation to be scaled orthogonally to existing sequential methods. Our work makes two key innovations: First, we present A2R as a plug-and-play parallel reasoning framework that explicitly enhances a model's capabilities on complex questions. For example, using our framework, the Qwen3-8B-distill model achieves a 75% performance improvement compared to its self-consistency baseline. Second, through a systematic analysis of the explorer and synthesizer roles, we identify an effective asymmetric scaling paradigm. This insight leads to A2R-Efficient, a "small-to-big" variant that combines a Qwen3-4B explorer with a Qwen3-8B synthesizer. This configuration surpasses the average performance of a monolithic Qwen3-32B model at a nearly 30% lower cost. Collectively, these results show that A2R is not only a performance-boosting framework but also an efficient and practical solution for real-world applications.