From Insight to Action: A Novel Framework for Interpretability-Guided Data Selection in Large Language Models

While mechanistic interpretability tools like Sparse Autoencoders (SAEs) can uncover meaningful features within Large Language Models (LLMs), a critical gap remains in transforming these insights into practical actions for model optimization. We bridge this gap with the hypothesis that data selection guided by a model's internal task features is a effective training strategy. Inspired by this, we propose Interpretability-Guided Data Selection (IGDS), a framework that first identifies these causal task features through frequency recall and interventional filtering, then selects ``Feature-Resonant Data'' that maximally activates task features for fine-tuning. We validate IGDS on mathematical reasoning, summarization, and translation tasks within Gemma-2, LLaMA-3.1, and Qwen3 models. Our experiments demonstrate exceptional data efficiency: on the Math task, IGDS surpasses full-dataset fine-tuning by a remarkable 17.4% on Gemma-2-2B while using only 50% of the data, and outperforms established baselines focused on data quality and diversity. Analysis confirms a strong positive correlation between feature amplification and task performance improvement. IGDS thus provides a direct and effective framework to enhance LLMs by leveraging their internal mechanisms, validating our core hypothesis.

X2-N: A Transformable Wheel-legged Humanoid Robot with Dual-mode Locomotion and Manipulation

Wheel-legged robots combine the efficiency of wheeled locomotion with the versatility of legged systems, enabling rapid traversal over both continuous and discrete terrains. However, conventional designs typically employ fixed wheels as feet and limited degrees of freedom (DoFs) at the hips, resulting in reduced stability and mobility during legged locomotion compared to humanoids with flat feet. In addition, most existing platforms lack a full upper body with arms, which limits their ability to perform dexterous manipulation tasks. In this letter, we present X2-N, a high-DoF transformable robot with dual-mode locomotion and manipulation. X2-N can operate in both humanoid and wheel-legged forms and transform seamlessly between them through joint reconfiguration. We further propose a reinforcement learning (RL)-based whole-body control framework tailored to this morphology, enabling unified control across hybrid locomotion, transformation, and manipulation. We validate X2-N in a range of challenging locomotion and manipulation tasks, including dynamic skating-like motion, stair climbing and package delivery. Results demonstrate high locomotion efficiency, strong terrain adaptability, and stable loco-manipulation performance of X2-N, highlighting its potential for real-world deployment.

R2IF: Aligning Reasoning with Decisions via Composite Rewards for Interpretable LLM Function Calling

Function calling empowers large language models (LLMs) to interface with external tools, yet existing RL-based approaches suffer from misalignment between reasoning processes and tool-call decisions. We propose R2IF, a reasoning-aware RL framework for interpretable function calling, adopting a composite reward integrating format/correctness constraints, Chain-of-Thought Effectiveness Reward (CER), and Specification-Modification-Value (SMV) reward, optimized via GRPO. Experiments on BFCL/ACEBench show R2IF outperforms baselines by up to 34.62% (Llama3.2-3B on BFCL) with positive Average CoT Effectiveness (0.05 for Llama3.2-3B), enhancing both function-calling accuracy and interpretability for reliable tool-augmented LLM deployment.

SAGE: A Service Agent Graph-guided Evaluation Benchmark

The development of Large Language Models (LLMs) has catalyzed automation in customer service, yet benchmarking their performance remains challenging. Existing benchmarks predominantly rely on static paradigms and single-dimensional metrics, failing to account for diverse user behaviors or the strict adherence to structured Standard Operating Procedures (SOPs) required in real-world deployments. To bridge this gap, we propose SAGE (Service Agent Graph-guided Evaluation), a universal multi-agent benchmark for automated, dual-axis assessment. SAGE formalizes unstructured SOPs into Dynamic Dialogue Graphs, enabling precise verification of logical compliance and comprehensive path coverage. We introduce an Adversarial Intent Taxonomy and a modular Extension Mechanism, enabling low-cost deployment across domains and facilitating automated dialogue data synthesis. Evaluation is conducted via a framework where Judge Agents and a Rule Engine analyze interactions between User and Service Agents to generate deterministic ground truth. Extensive experiments on 27 LLMs across 6 industrial scenarios reveal a significant ``Execution Gap'' where models accurately classify intents but fail to derive correct subsequent actions. We also observe ``Empathy Resilience'', a phenomenon where models maintain polite conversational facades despite underlying logical failures under high adversarial intensity. Code and resources are available at https://anonymous.4open.science/r/SAGE-Bench-4CD3/.

DEP: A Decentralized Large Language Model Evaluation Protocol

With the rapid development of Large Language Models (LLMs), a large number of benchmarks have been proposed. However, most benchmarks lack unified evaluation standard and require the manual implementation of custom scripts, making results hard to ensure consistency and reproducibility. Furthermore, mainstream evaluation frameworks are centralized, with datasets and answers, which increases the risk of benchmark leakage. To address these issues, we propose a Decentralized Evaluation Protocol (DEP), a decentralized yet unified and standardized evaluation framework through a matching server without constraining benchmarks. The server can be mounted locally or deployed remotely, and once adapted, it can be reused over the long term. By decoupling users, LLMs, and benchmarks, DEP enables modular, plug-and-play evaluation: benchmark files and evaluation logic stay exclusively on the server side. In remote setting, users cannot access the ground truth, thereby achieving data isolation and leak-proof evaluation. To facilitate practical adoption, we develop DEP Toolkit, a protocol-compatible toolkit that supports features such as breakpoint resume, concurrent requests, and congestion control. We also provide detailed documentation for adapting new benchmarks to DEP. Using DEP toolkit, we evaluate multiple LLMs across benchmarks. Experimental results verify the effectiveness of DEP and show that it reduces the cost of deploying benchmark evaluations. As of February 2026, we have adapted over 60 benchmarks and continue to promote community co-construction to support unified evaluation across various tasks and domains.

Reinforcing Real-world Service Agents: Balancing Utility and Cost in Task-oriented Dialogue

The rapid evolution of Large Language Models (LLMs) has accelerated the transition from conversational chatbots to general agents. However, effectively balancing empathetic communication with budget-aware decision-making remains an open challenge. Since existing methods fail to capture these complex strategic trade-offs, we propose InteractCS-RL, a framework that reframes task-oriented dialogue as a multi-granularity reinforcement learning process. Specifically, we first establish a User-centric Interaction Framework to provide a high-fidelity training gym, enabling agents to dynamically explore diverse strategies with persona-driven users. Then, we introduce Cost-aware Multi-turn Policy Optimization (CMPO) with a hybrid advantage estimation strategy. By integrating generative process credits and employing a PID-Lagrangian cost controller, CMPO effectively guides the policy to explore Pareto boundary between user reward and global cost constraints. Extensive experiments on customized real business scenarios demonstrate that InteractCS-RL significantly outperform other baselines across three evaluation dimensions. Further evaluation on tool-agent-user interaction benchmarks verify InteractCS-RL robustness across diverse domains.

A Fast Method for Planning All Optimal Homotopic Configurations for Tethered Robots and Its Extended Applications

Tethered robots play a pivotal role in specialized environments such as disaster response and underground exploration, where their stable power supply and reliable communication offer unparalleled advantages. However, their motion planning is severely constrained by tether length limitations and entanglement risks, posing significant challenges to achieving optimal path planning. To address these challenges, this study introduces CDT-TCS (Convex Dissection Topology-based Tethered Configuration Search), a novel algorithm that leverages CDT Encoding as a homotopy invariant to represent topological states of paths. By integrating algebraic topology with geometric optimization, CDT-TCS efficiently computes the complete set of optimal feasible configurations for tethered robots at all positions in 2D environments through a single computation. Building on this foundation, we further propose three application-specific algorithms: i) CDT-TPP for optimal tethered path planning, ii) CDT-TMV for multi-goal visiting with tether constraints, iii) CDT-UTPP for distance-optimal path planning of untethered robots. All theoretical results and propositions underlying these algorithms are rigorously proven and thoroughly discussed in this paper. Extensive simulations demonstrate that the proposed algorithms significantly outperform state-of-the-art methods in their respective problem domains. Furthermore, real-world experiments on robotic platforms validate the practicality and engineering value of the proposed framework.

PierGuard: A Planning Framework for Underwater Robotic Inspection of Coastal Piers

Using underwater robots instead of humans for the inspection of coastal piers can enhance efficiency while reducing risks. A key challenge in performing these tasks lies in achieving efficient and rapid path planning within complex environments. Sampling-based path planning methods, such as Rapidly-exploring Random Tree* (RRT*), have demonstrated notable performance in high-dimensional spaces. In recent years, researchers have begun designing various geometry-inspired heuristics and neural network-driven heuristics to further enhance the effectiveness of RRT*. However, the performance of these general path planning methods still requires improvement when applied to highly cluttered underwater environments. In this paper, we propose PierGuard, which combines the strengths of bidirectional search and neural network-driven heuristic regions. We design a specialized neural network to generate high-quality heuristic regions in cluttered maps, thereby improving the performance of the path planning. Through extensive simulation and real-world ocean field experiments, we demonstrate the effectiveness and efficiency of our proposed method compared with previous research. Our method achieves approximately 2.6 times the performance of the state-of-the-art geometric-based sampling method and nearly 4.9 times that of the state-of-the-art learning-based sampling method. Our results provide valuable insights for the automation of pier inspection and the enhancement of maritime safety. The updated experimental video is available in the supplementary materials.

ProBench: Benchmarking Large Language Models in Competitive Programming

With reasoning language models such as OpenAI-o3 and DeepSeek-R1 emerging, large language models (LLMs) have entered a new phase of development. However, existing benchmarks for coding evaluation are gradually inadequate to assess the capability of advanced LLMs in code reasoning. To bridge the gap for high-level code reasoning assessment, we propose ProBench to benchmark LLMs in competitive programming, drawing inspiration from the International Collegiate Programming Contest. ProBench collects a comprehensive set of competitive programming problems from Codeforces, Luogu, and Nowcoder platforms during the period from July to December 2024, obtaining real test results through online submissions to ensure the fairness and accuracy of the evaluation. We establish a unified problem attribute system, including difficulty grading and algorithm tagging. With carefully collected and annotated data in ProBench, we systematically assess 9 latest LLMs in competitive programming across multiple dimensions, including thought chain analysis, error type diagnosis, and reasoning depth evaluation. Experimental results show that QwQ-32B-Preview achieves the best score of 20.93 followed by DeepSeek-V3 with a score of 16.38, suggesting that models trained with specialized reasoning tasks significantly outperform general-purpose models (even larger than reasoning-oriented models) in programming. Further analysis also reveals key areas for programming capability enhancement, e.g., algorithm adaptability and reasoning sufficiency, providing important insights for the future development of reasoning models.

Detecting LLM Fact-conflicting Hallucinations Enhanced by Temporal-logic-based Reasoning

Large language models (LLMs) face the challenge of hallucinations -- outputs that seem coherent but are actually incorrect. A particularly damaging type is fact-conflicting hallucination (FCH), where generated content contradicts established facts. Addressing FCH presents three main challenges: 1) Automatically constructing and maintaining large-scale benchmark datasets is difficult and resource-intensive; 2) Generating complex and efficient test cases that the LLM has not been trained on -- especially those involving intricate temporal features -- is challenging, yet crucial for eliciting hallucinations; and 3) Validating the reasoning behind LLM outputs is inherently difficult, particularly with complex logical relationships, as it requires transparency in the model's decision-making process. This paper presents Drowzee, an innovative end-to-end metamorphic testing framework that utilizes temporal logic to identify fact-conflicting hallucinations (FCH) in large language models (LLMs). Drowzee builds a comprehensive factual knowledge base by crawling sources like Wikipedia and uses automated temporal-logic reasoning to convert this knowledge into a large, extensible set of test cases with ground truth answers. LLMs are tested using these cases through template-based prompts, which require them to generate both answers and reasoning steps. To validate the reasoning, we propose two semantic-aware oracles that compare the semantic structure of LLM outputs to the ground truths. Across nine LLMs in nine different knowledge domains, experimental results show that Drowzee effectively identifies rates of non-temporal-related hallucinations ranging from 24.7% to 59.8%, and rates of temporal-related hallucinations ranging from 16.7% to 39.2%.