Large Language Model (LLM) unlearning is essential for removing sensitive or copyrighted knowledge while preserving general utility. Existing methods often leave residual knowledge in intermediate representations, which can still be recovered. To address this, we propose Cascade, a hierarchical recoverability control framework that minimizes the internal identifiability of target knowledge. Cascade combines three complementary controls: path-level routing to suppress privacy-associated activation routes, representation-level compression to reduce geometric separability, and decoding-level intervention to limit residual recovery. Experiments on TOFU, MUSE-News, and WMDP, including robustness tests with query reformulation and extraction-style prompts, show that Cascade effectively reduces recoverability while maintaining stable model utility.

Hidden states of large language models (LLMs) are typically analyzed as flat Euclidean vectors through probing, anisotropy, and intrinsic-dimension lenses. Yet much of what these models process (taxonomies, program structure, multi-step derivations) is hierarchical, and whether hidden-state distances reflect such hierarchy remains untested. We present the first systematic study of Gromov hyperbolicity (GH), a distance-based measure of tree-likeness, in prompt-token hidden states of open-weight LLMs. Across model families, scales, layer depths, and domains spanning math, code, commonsense, and truthfulness benchmarks, we construct a hyperbolicity atlas charting the normalized statistic 𝛿_rel along each axis. Four findings emerge: (i) depth dominates, with middle layers forming a high-𝛿_rel plateau and final layers dropping toward tree-like structure; (ii) scale is weak and non-monotonic; (iii) matched-scale families occupy distinct regimes, with specialization gaps exceeding scale effects; and (iv) domains interact with specialization, e.g., HumanEval shifts from the highest- to the lowest-𝛿_rel domain between base and coder checkpoints. GH thus serves as a lightweight diagnostic that localizes where hierarchical distance structure emerges and which comparisons warrant deeper analysis.

With the release of OpenAI’s o1 model, reasoning models that adopt slow-thinking strategies have become increasingly common. Their outputs often contain complex reasoning, intermediate steps, and self-reflection, making existing evaluation methods and reward models inadequate. In particular, they struggle to judge answer equivalence and to reliably extract final answers from long, complex responses. To address this challenge, we propose xVerify, an efficient answer verifier for evaluating reasoning models. xVerify shows strong equivalence judgment capabilities, enabling accurate comparison between model outputs and reference answers across diverse question types. To train and evaluate xVerify, we construct the VAR dataset, consisting of multi-round annotated QA pairs generated by multiple LLMs across challenging reasoning tasks. Experimental results on both test and generalization sets show that all xVerify variants achieve over 95% F1 score and accuracy. Notably, the smallest model, xVerify-0.5B-I, outperforms all evaluation methods except GPT-4o, while xVerify-3B-Ib surpasses GPT-4o in overall performance. In addition, reinforcement learning experiments using xVerify as the reward model yield an 18.4% improvement for Qwen2.5-7B compared with direct generation, exceeding the gains achieved with Math Verify as the reward. These results demonstrate the effectiveness and generalizability of xVerify. All xVerify resources are available on [GitHub](https://github.com/IAAR-Shanghai/xVerify).

Client-side privacy rewriting is crucial for deploying LLMs in privacy-sensitive domains. However, existing approaches struggle to balance privacy and utility. Full-text methods often distort context, while span-level approaches rely on impractical manual masks or brittle static dictionaries. Attempts to automate localization via prompt-based LLMs prove unreliable, as they suffer from unstable instruction following that leads to privacy leakage and excessive context scrubbing. To address these limitations, we propose DAMPER (Domain-Aware Mask-free Privacy Extraction and Rewriting). DAMPER operationalizes latent privacy semantics into compact Domain Privacy Prototypes via contrastive learning, enabling precise, autonomous span localization. Furthermore, we introduce a Prototype-Guided Preference Alignment, which leverages learned prototypes as semantic anchors to construct preference pairs, optimizing a domain-compliant rewriting policy without human annotations. At inference time, DAMPER integrates a sampling-based Exponential Mechanism to provide rigorous span-level Differential Privacy (DP) guarantees. Extensive experiments demonstrate that DAMPER significantly outperforms existing baselines, achieving a superior privacy-utility trade-off.

With rapid advancements in large language models (LLMs) across artificial intelligence, machine learning, and data sci-ence, there is a growing need for evaluation frameworks that go beyond traditional performance metrics. Conventional methods focus mainly on accuracy and computational metrics, often neglecting user experience and community interaction—key elements in open-source environments. This paper intro-duces a multi-dimensional, user-centered evaluation frame-work, integrating metrics like User Engagement Index (UEI), Community Response Rate (CRR), and a Time Weight Factor (TWF) to assess LLMs’ real-world impact. Additionally, we propose an adaptive weighting mechanism using Bayesian op-timization to dynamically adjust metric weights for more ac-curate model evaluation. Experimental results confirm that our framework effectively identifies models with strong user engagement and community support, offering a balanced, data-driven approach to open-source LLM evaluation. This frame-work serves as a valuable tool for developers and researchers in selecting and improving open-source models.

The evaluation of large language models (LLMs) has traditionally relied on static benchmarks, a paradigm that poses two major limitations: (1) predefined test sets lack adaptability to diverse application domains, and (2) standardized evaluation protocols often fail to capture fine-grained assessments of domain-specific knowledge and contextual reasoning abilities. To overcome these challenges, we propose GuessArena, an adaptive evaluation framework grounded in adversarial game-based interactions. Inspired by the interactive structure of the Guess Who I Am? game, our framework seamlessly integrates dynamic domain knowledge modeling with progressive reasoning assessment to improve evaluation fidelity. Empirical studies across five vertical domains-finance, healthcare, manufacturing, information technology, and education-demonstrate that GuessArena effectively distinguishes LLMs in terms of domain knowledge coverage and reasoning chain completeness. Compared to conventional benchmarks, our method provides substantial advantages in interpretability, scalability, and scenario adaptability.