Large language models have transformed code generation, raising concerns around authorship, assessment integrity, and software trust. SemEval-2026 Task 13 Subtask A operationalizes detection as binary classification over code snippets, with a particular emphasis on out-of-distribution (OOD) generalization across unseen programming languages and application domains. We propose a SALSA-style formulation, Single-pass Autoregressive LLM Structured Classification, that maps each class to a dedicated output token and trains the model to emit a single-token label in a structured response. Rather than engineering hand-crafted features or decision rules, this formulation delegates the authorship decision to the model. To improve OOD robustness, we combine balanced sampling across languages with parameter-efficient fine-tuning and conservative training (low learning rate, single epoch) to avoid overfitting to the training domain. Our best system achieves OOD $F_1 = 0.789$ on the official leaderboard, substantially outperforming the CodeBERT baseline ($F_1 = 0.305$).
展开完整摘要收起摘要↓
Large language models have transformed code generation, raising concerns around authorship, assessment integrity, and software trust. SemEval-2026 Task 13 Subtask A operationalizes detection as binary classification over code snippets, with a particular emphasis on out-of-distribution (OOD) generalization across unseen programming languages and application domains. We propose a SALSA-style formulation, Single-pass Autoregressive LLM Structured Classification, that maps each class to a dedicated output token and trains the model to emit a single-token label in a structured response. Rather than engineering hand-crafted features or decision rules, this formulation delegates the authorship decision to the model. To improve OOD robustness, we combine balanced sampling across languages with parameter-efficient fine-tuning and conservative training (low learning rate, single epoch) to avoid overfitting to the training domain. Our best system achieves OOD $F_1 = 0.789$ on the official leaderboard, substantially outperforming the CodeBERT baseline ($F_1 = 0.305$).
作者Cosimo Galeone, Anna Ettorre, Minsu Park, Giuseppe Ettorre, Daniele Ligorio
A central aspiration of mechanistic interpretability is controllability: if we know where a behavior is represented in a model's activations, we should be able to modify it. This rests on a hidden premise — that the direction which detects a behavior and the direction which controls it are the same, or close. We test this geometrically: what is the angle between the direction that best detects a behavior and the one that best causes it? If detection implies control the cosine is near 1; otherwise it quantifies a detection-intervention gap. On Gemma 2-2B-it, output format (clean JSON vs markdown fencing) collapses both roles onto one axis. Hallucination does not: the model detects fake entities with perfect linear separability (AUC = 1.000 from layer 5), yet that direction sits at cos = 0.12 (about 83 degrees) from the direction producing a refusal — a small, reproducible alignment, far from the cos = 1 that "detection is control" would require. A detector built from activations, with no chosen tokens, likewise fails to align (cos = -0.06). The gap generalizes: across four models from three families and two scales (1B-9B), cos stays in [0.12, 0.20], identical before and after instruction tuning (0.1197 vs 0.1200), placing its origin in pretraining. A 15-degree rotation toward the refusal direction partially bridges it — 73% and 60% refusal on two held-out fake-entity categories at 1.8% false positives. We then ask whether this cosine predicts steerability, and it does not: detection is a high-dimensional class, not a single direction, and what separates the steerable case is functional, not readable from a static angle. The cosine is a weight-computable signature of the dissociation between knowing and steering, not a predictor of it.
展开完整摘要收起摘要↓
A central aspiration of mechanistic interpretability is controllability: if we know where a behavior is represented in a model's activations, we should be able to modify it. This rests on a hidden premise — that the direction which detects a behavior and the direction which controls it are the same, or close. We test this geometrically: what is the angle between the direction that best detects a behavior and the one that best causes it? If detection implies control the cosine is near 1; otherwise it quantifies a detection-intervention gap. On Gemma 2-2B-it, output format (clean JSON vs markdown fencing) collapses both roles onto one axis. Hallucination does not: the model detects fake entities with perfect linear separability (AUC = 1.000 from layer 5), yet that direction sits at cos = 0.12 (about 83 degrees) from the direction producing a refusal — a small, reproducible alignment, far from the cos = 1 that "detection is control" would require. A detector built from activations, with no chosen tokens, likewise fails to align (cos = -0.06). The gap generalizes: across four models from three families and two scales (1B-9B), cos stays in [0.12, 0.20], identical before and after instruction tuning (0.1197 vs 0.1200), placing its origin in pretraining. A 15-degree rotation toward the refusal direction partially bridges it — 73% and 60% refusal on two held-out fake-entity categories at 1.8% false positives. We then ask whether this cosine predicts steerability, and it does not: detection is a high-dimensional class, not a single direction, and what separates the steerable case is functional, not readable from a static angle. The cosine is a weight-computable signature of the dissociation between knowing and steering, not a predictor of it.
作者Pietro Tropeano, Maria Maistro, Tuukka Ruotsalo, Christina Lioma
Pruning Large Language Models (LLMs) reduces memory and inference costs by removing parts of the network, producing smaller models that retain most of their accuracy. As attention layers are the most resource-intensive parts of LLMs, pruning them is a promising compression strategy. Prior work shows that up to 33% of attention layers can be pruned with minimal accuracy loss. Nevertheless, the impact of attention pruning on model interpretability, specifically faithfulness and confidence calibration, remains unstudied. To address this gap, we study how pruning attention layers affects explanation faithfulness and confidence calibration across five LLMs and eight datasets. While the pruned models often maintain high accuracy, we find that their faithfulness and calibration often degrade. Notably, faithfulness and calibration can fluctuate significantly, even when accuracy remains stable, highlighting a misalignment between model confidence, interpretability, and accuracy. Our findings suggest that layer pruning can affect LLMs' interpretability and reliability in ways not captured by accuracy and efficiency measures alone. We recommend including explainability and calibration metrics when evaluating pruned models.
展开完整摘要收起摘要↓
Pruning Large Language Models (LLMs) reduces memory and inference costs by removing parts of the network, producing smaller models that retain most of their accuracy. As attention layers are the most resource-intensive parts of LLMs, pruning them is a promising compression strategy. Prior work shows that up to 33% of attention layers can be pruned with minimal accuracy loss. Nevertheless, the impact of attention pruning on model interpretability, specifically faithfulness and confidence calibration, remains unstudied. To address this gap, we study how pruning attention layers affects explanation faithfulness and confidence calibration across five LLMs and eight datasets. While the pruned models often maintain high accuracy, we find that their faithfulness and calibration often degrade. Notably, faithfulness and calibration can fluctuate significantly, even when accuracy remains stable, highlighting a misalignment between model confidence, interpretability, and accuracy. Our findings suggest that layer pruning can affect LLMs' interpretability and reliability in ways not captured by accuracy and efficiency measures alone. We recommend including explainability and calibration metrics when evaluating pruned models.
作者Vatsal Baherwani, Zixi Chen, Shikai Qiu, Andrew Gordon Wilson, Pavel Izmailov
Neural scaling laws for transformer language models predict smooth improvements in pretraining loss with increasing parameters, but downstream capabilities such as in-context learning are known to emerge abruptly past a certain model scale. In this paper, we show that emergent capabilities arise stochastically throughout training, with larger models acquiring them earlier on average. We demonstrate that the emergence of capabilities such as pattern completion and indirect object identification corresponds to the abrupt learning of task-relevant attention patterns. To isolate this phenomenon, we train transformer models on synthetic linear map and cellular automata datasets, and we show that the difficulty of learning attention patterns depends on context length and pattern sparsity. Moreover, scaling the number of attention heads improves learning efficiency on our synthetic tasks, while increasing the head dimension yields diminishing returns past a minimum capacity. We additionally investigate architectures with alternative attention mechanisms, showing that MLP-Mixer outperforms a transformer on linear map tasks with complex attention patterns. Our findings provide a mechanistic insight into emergence, showing that downstream capabilities arise abruptly due to the intrinsic difficulty of learning sparse attention patterns in transformer models.
展开完整摘要收起摘要↓
Neural scaling laws for transformer language models predict smooth improvements in pretraining loss with increasing parameters, but downstream capabilities such as in-context learning are known to emerge abruptly past a certain model scale. In this paper, we show that emergent capabilities arise stochastically throughout training, with larger models acquiring them earlier on average. We demonstrate that the emergence of capabilities such as pattern completion and indirect object identification corresponds to the abrupt learning of task-relevant attention patterns. To isolate this phenomenon, we train transformer models on synthetic linear map and cellular automata datasets, and we show that the difficulty of learning attention patterns depends on context length and pattern sparsity. Moreover, scaling the number of attention heads improves learning efficiency on our synthetic tasks, while increasing the head dimension yields diminishing returns past a minimum capacity. We additionally investigate architectures with alternative attention mechanisms, showing that MLP-Mixer outperforms a transformer on linear map tasks with complex attention patterns. Our findings provide a mechanistic insight into emergence, showing that downstream capabilities arise abruptly due to the intrinsic difficulty of learning sparse attention patterns in transformer models.
Recovering governing Ordinary Differential Equations (ODEs) from data is a central challenge in modeling dynamical systems across scientific domains. Existing approaches cast discovery as a static inference problem over fixed datasets, assuming that the observed trajectories are sufficiently informative. However, dynamical systems evolve over large state spaces, and limited data can make multiple equations observationally indistinguishable, leading to identifiability gaps and the recovery of incorrect governing equations. To address this, we introduce LLM-ACES, or LLM-guided Active Closed-loop Equation Search, a closed-loop framework that jointly optimizes symbolic hypothesis construction and adaptive data acquisition. In LLM-ACES, a large language model (LLM) proposes operator priors that partition the large search space into distinct regions, within which candidate equations are fit to the observed data. The disagreement among these candidates guides the acquisition of informative trajectories, creating a feedback loop that iteratively refines both the hypothesis space and the discovered dynamics. On 122 ODE systems spanning ODEBench and ODEBase, LLM-ACES achieves the lowest median NMSE, outperforming state-of-the-art baselines by several orders of magnitude while achieving a high symbolic accuracy of 46.2% and 52.4%, respectively. Our analysis further shows that LLM-ACES is sample-efficient, achieving better performance with one-tenth the data. Furthermore, LLM-ACES's feedback-driven data acquisition makes it robust to noise and recovers the correct symbolic structure, while baselines introduce spurious terms that fit the data locally but obscure the true governing relationships.
展开完整摘要收起摘要↓
Recovering governing Ordinary Differential Equations (ODEs) from data is a central challenge in modeling dynamical systems across scientific domains. Existing approaches cast discovery as a static inference problem over fixed datasets, assuming that the observed trajectories are sufficiently informative. However, dynamical systems evolve over large state spaces, and limited data can make multiple equations observationally indistinguishable, leading to identifiability gaps and the recovery of incorrect governing equations. To address this, we introduce LLM-ACES, or LLM-guided Active Closed-loop Equation Search, a closed-loop framework that jointly optimizes symbolic hypothesis construction and adaptive data acquisition. In LLM-ACES, a large language model (LLM) proposes operator priors that partition the large search space into distinct regions, within which candidate equations are fit to the observed data. The disagreement among these candidates guides the acquisition of informative trajectories, creating a feedback loop that iteratively refines both the hypothesis space and the discovered dynamics. On 122 ODE systems spanning ODEBench and ODEBase, LLM-ACES achieves the lowest median NMSE, outperforming state-of-the-art baselines by several orders of magnitude while achieving a high symbolic accuracy of 46.2% and 52.4%, respectively. Our analysis further shows that LLM-ACES is sample-efficient, achieving better performance with one-tenth the data. Furthermore, LLM-ACES's feedback-driven data acquisition makes it robust to noise and recovers the correct symbolic structure, while baselines introduce spurious terms that fit the data locally but obscure the true governing relationships.
作者Kun Zhang, Chenxin Fang, Tao Chen, Baiyang Song, Yunhang Shen, Yiyi Zhou, Rongrong Ji
Long video understanding remains a daunting challenge for Multimodal Large Language Models (MLLMs) due to the excessive computation and memory footprint. Thus, keyframe selection is often adopted to mitigate this shortcoming, which however still suffers from low flexibility and high noise due to its hard sampling principle. In this paper, we define video frame selection as a problem of Quasi-Gaussian Sampling, and propose an adaptive and training-free approach termed AdaQ. Inspired by the 3-$σ$ rule of Gaussian distribution, the objective of AdaQ is to achieve the optimal 3-$σ$ interval for different examples, i.e., a smaller 3-$σ$ interval for the local query and a larger one for the global query, thereby facilitating robust and adaptive frame sampling. To validate AdaQ, we apply it to four MLLMs with three embedding models. The extensive experimental results not only show its obvious performance gains over the default MLLMs and the SOTA keyframe selection methods, e.g., helping Qwen3-VL-8B outperform GPT4o by 15.8% on average by using only 64 frames, but also confirm its superior robustness and high efficiency for long-video understanding, e.g., only 1 hyper-parameter needs to be set.
展开完整摘要收起摘要↓
Long video understanding remains a daunting challenge for Multimodal Large Language Models (MLLMs) due to the excessive computation and memory footprint. Thus, keyframe selection is often adopted to mitigate this shortcoming, which however still suffers from low flexibility and high noise due to its hard sampling principle. In this paper, we define video frame selection as a problem of Quasi-Gaussian Sampling, and propose an adaptive and training-free approach termed AdaQ. Inspired by the 3-$σ$ rule of Gaussian distribution, the objective of AdaQ is to achieve the optimal 3-$σ$ interval for different examples, i.e., a smaller 3-$σ$ interval for the local query and a larger one for the global query, thereby facilitating robust and adaptive frame sampling. To validate AdaQ, we apply it to four MLLMs with three embedding models. The extensive experimental results not only show its obvious performance gains over the default MLLMs and the SOTA keyframe selection methods, e.g., helping Qwen3-VL-8B outperform GPT4o by 15.8% on average by using only 64 frames, but also confirm its superior robustness and high efficiency for long-video understanding, e.g., only 1 hyper-parameter needs to be set.
Large language model (LLM) agents increasingly automate complex tasks by integrating language models with external tools and environments. However, their autonomy poses significant safety risks: agents may execute destructive commands, leak sensitive data, or violate domain constraints. Existing safety approaches face a fundamental tradeoff: hand-crafted rules are interpretable but brittle, with overly conservative rules blocking safe operations (high false positives) while permissive rules miss unsafe behaviors (high false negatives). Neural classifiers lack the interpretability required for safety-critical deployments. We present AutoSpec, a framework that automatically evolves deployed expert-designed safety rules from user safe/unsafe annotations through counterexample-guided inductive synthesis (CEGIS) guided by inductive logic programming (ILP). Starting from the expert rules and a stream of annotated traces, AutoSpec iteratively evaluates rules, mines false-positive and false-negative counterexamples, uses ILP to learn which predicates discriminate them, generates candidate rule edits, and verifies candidates to select the best revision. The key insight is that ILP efficiently identifies predicates that appear frequently in false negatives but rarely in false positives (or vice versa), dramatically pruning the exponential search space of rule edits. This continues until convergence, producing interpretable rules that balance precision and recall. We evaluate AutoSpec on 291 execution traces spanning code execution and embodied agent domains. AutoSpec raises rule F1 to 0.98 and 0.93 across the two domains, achieving up to 94% false positive reduction while maintaining high recall, and converges within 4-5 iterations. The ILP-guided approach achieves up to 4.8x higher F1 than heuristic CEGIS. The learned rules are human-readable, auditable, and generalize to unseen scenarios.
展开完整摘要收起摘要↓
Large language model (LLM) agents increasingly automate complex tasks by integrating language models with external tools and environments. However, their autonomy poses significant safety risks: agents may execute destructive commands, leak sensitive data, or violate domain constraints. Existing safety approaches face a fundamental tradeoff: hand-crafted rules are interpretable but brittle, with overly conservative rules blocking safe operations (high false positives) while permissive rules miss unsafe behaviors (high false negatives). Neural classifiers lack the interpretability required for safety-critical deployments. We present AutoSpec, a framework that automatically evolves deployed expert-designed safety rules from user safe/unsafe annotations through counterexample-guided inductive synthesis (CEGIS) guided by inductive logic programming (ILP). Starting from the expert rules and a stream of annotated traces, AutoSpec iteratively evaluates rules, mines false-positive and false-negative counterexamples, uses ILP to learn which predicates discriminate them, generates candidate rule edits, and verifies candidates to select the best revision. The key insight is that ILP efficiently identifies predicates that appear frequently in false negatives but rarely in false positives (or vice versa), dramatically pruning the exponential search space of rule edits. This continues until convergence, producing interpretable rules that balance precision and recall. We evaluate AutoSpec on 291 execution traces spanning code execution and embodied agent domains. AutoSpec raises rule F1 to 0.98 and 0.93 across the two domains, achieving up to 94% false positive reduction while maintaining high recall, and converges within 4-5 iterations. The ILP-guided approach achieves up to 4.8x higher F1 than heuristic CEGIS. The learned rules are human-readable, auditable, and generalize to unseen scenarios.
作者Jessica Chudnovsky, Joshua Kazdan, Noam Levi, Rylan Schaeffer, Yegor Denisov-Blanch, Bo He, Mehmet Donmez, Sanmi Koyejo, David Donoho
Language models are running out of high-quality training data, and even aggressively deduplicated corpora retain some amount of repetition. Earlier controlled studies predated Chinchilla-style scaling laws and could only measure the cost of repetition indirectly. We revisit repetition in the Chinchilla era, using a fitted no-repetition scaling law to report Compute-Equivalent Gain and Compute-Equivalent Loss. We show that under this modernized paradigm, repetition damage is systematic in three ways. First, holding compute allocated to repeated data constant, eval loss peaks at an intermediate repeat count $\Rep$; repeating a moderately sized subset a moderate number of times damages performance more than repeating a large subset a few times or a small subset many times. Second, the location of this peak is well-fit by a power law in model size; this scaling law reveals that the most damaging number of repeated data grows more quickly than compute. Finally, when repeated documents consume 10% of the FLOPs budget in a controlled exact-document repetition setting, the compute-equivalent loss can be large: on FineWeb-Edu-Dedup, the most damaging repeat count for a Qwen3-style 344M-parameter model at $\OT=1$ matches the loss of a no-repetition run using 67% of the FLOPs. We demonstrate that these phenomena are not language-model-specific, and can be analytically understood in a simple statistical model: a misspecified linear regression with verbatim duplicates reproduces the same qualitative loss peak, quantifying how such peaks can arise from a statistical tradeoff between memorization and generalization. Our findings add precision to the study of duplication in language models, allowing practitioners to quantify the wasted compute incurred by the presence and repeat structure of duplicates in pretraining corpora.
展开完整摘要收起摘要↓
Language models are running out of high-quality training data, and even aggressively deduplicated corpora retain some amount of repetition. Earlier controlled studies predated Chinchilla-style scaling laws and could only measure the cost of repetition indirectly. We revisit repetition in the Chinchilla era, using a fitted no-repetition scaling law to report Compute-Equivalent Gain and Compute-Equivalent Loss. We show that under this modernized paradigm, repetition damage is systematic in three ways. First, holding compute allocated to repeated data constant, eval loss peaks at an intermediate repeat count $\Rep$; repeating a moderately sized subset a moderate number of times damages performance more than repeating a large subset a few times or a small subset many times. Second, the location of this peak is well-fit by a power law in model size; this scaling law reveals that the most damaging number of repeated data grows more quickly than compute. Finally, when repeated documents consume 10% of the FLOPs budget in a controlled exact-document repetition setting, the compute-equivalent loss can be large: on FineWeb-Edu-Dedup, the most damaging repeat count for a Qwen3-style 344M-parameter model at $\OT=1$ matches the loss of a no-repetition run using 67% of the FLOPs. We demonstrate that these phenomena are not language-model-specific, and can be analytically understood in a simple statistical model: a misspecified linear regression with verbatim duplicates reproduces the same qualitative loss peak, quantifying how such peaks can arise from a statistical tradeoff between memorization and generalization. Our findings add precision to the study of duplication in language models, allowing practitioners to quantify the wasted compute incurred by the presence and repeat structure of duplicates in pretraining corpora.
Large language model (LLM) agents rely on long-term memory to support extended interactions and personalized assistance beyond finite context windows. Existing memory agents actively update external memory through generated write, revise, and delete operations, but these updates may omit important information, corrupt existing memory, or introduce unsupported hallucinated content. Once stored, such errors become persistent system-state failures that can affect future reasoning and generation. In this paper, we propose TrustMem, a framework designed to improve the trustworthiness of memory consolidation. TrustMem relies on a Memory Transition Verifier to evaluate the transition process of memory updates in terms of coverage, preservation, and faithfulness. It further constructs preference pairs among candidate updates under the same memory state, enabling preference-guided reinforcement learning to directly optimize memory updating behaviors. Extensive experiments demonstrate that TrustMem improves both memory utility and reliability: it achieves state-of-the-art results across MemoryAgentBench, HaluMem, and the Mem-alpha validation set, improves HaluMem memory extraction by 12.14 F1 points, and reduces transition-level omission, corruption, and hallucination by 40.1%, 79.1%, and 50.0%, respectively, compared with the strongest baseline for each error type.
展开完整摘要收起摘要↓
Large language model (LLM) agents rely on long-term memory to support extended interactions and personalized assistance beyond finite context windows. Existing memory agents actively update external memory through generated write, revise, and delete operations, but these updates may omit important information, corrupt existing memory, or introduce unsupported hallucinated content. Once stored, such errors become persistent system-state failures that can affect future reasoning and generation. In this paper, we propose TrustMem, a framework designed to improve the trustworthiness of memory consolidation. TrustMem relies on a Memory Transition Verifier to evaluate the transition process of memory updates in terms of coverage, preservation, and faithfulness. It further constructs preference pairs among candidate updates under the same memory state, enabling preference-guided reinforcement learning to directly optimize memory updating behaviors. Extensive experiments demonstrate that TrustMem improves both memory utility and reliability: it achieves state-of-the-art results across MemoryAgentBench, HaluMem, and the Mem-alpha validation set, improves HaluMem memory extraction by 12.14 F1 points, and reduces transition-level omission, corruption, and hallucination by 40.1%, 79.1%, and 50.0%, respectively, compared with the strongest baseline for each error type.
作者Gabriel Bravo-Palacios, Jianghan Zhang, Zachary Pestrikov, Brian Plancher, Thomas Lew
Robotics increasingly relies on GPUs for parallel simulation, large-scale learning, and neural-network inference. For model predictive control (MPC) to scale with this paradigm, solvers must run efficiently on this hardware while remaining fast, differentiable, and compatible with expressive MPC formulations used in robotics. We present TurboMPC, a differentiable MPC solver that runs entirely on the GPU and supports state and control inequality constraints, implicit integrators, cross-time-coupled costs, and slack variables. TurboMPC combines sequential quadratic programming (SQP), an alternating direction method of multipliers (ADMM) inner solver, implicit differentiation, and a co-designed JAX-CUDA implementation for efficiency and ease of use. In simulation, we validate TurboMPC on constrained planning, humanoid imitation learning, and reinforcement learning with neural-network cost function tasks, achieving up to $15\times$ and $58\times$ speedups over state-of-the-art CPU and GPU differentiable solvers, respectively. We deploy TurboMPC on a full-scale car for minimum-time racing and find that batched, GPU-accelerated tuning of MPC parameters via Bayesian optimization yields significantly faster driving than a hand-tuned baseline. TurboMPC also scales to planning horizons of over $8000$ knot points while maintaining control of the vehicle. We open-source TurboMPC at: https://github.com/ToyotaResearchInstitute/turbompc
展开完整摘要收起摘要↓
Robotics increasingly relies on GPUs for parallel simulation, large-scale learning, and neural-network inference. For model predictive control (MPC) to scale with this paradigm, solvers must run efficiently on this hardware while remaining fast, differentiable, and compatible with expressive MPC formulations used in robotics. We present TurboMPC, a differentiable MPC solver that runs entirely on the GPU and supports state and control inequality constraints, implicit integrators, cross-time-coupled costs, and slack variables. TurboMPC combines sequential quadratic programming (SQP), an alternating direction method of multipliers (ADMM) inner solver, implicit differentiation, and a co-designed JAX-CUDA implementation for efficiency and ease of use. In simulation, we validate TurboMPC on constrained planning, humanoid imitation learning, and reinforcement learning with neural-network cost function tasks, achieving up to $15\times$ and $58\times$ speedups over state-of-the-art CPU and GPU differentiable solvers, respectively. We deploy TurboMPC on a full-scale car for minimum-time racing and find that batched, GPU-accelerated tuning of MPC parameters via Bayesian optimization yields significantly faster driving than a hand-tuned baseline. TurboMPC also scales to planning horizons of over $8000$ knot points while maintaining control of the vehicle. We open-source TurboMPC at: https://github.com/ToyotaResearchInstitute/turbompc
MeMo proposes language models with explicit multi-layer correlation matrix memories (CMMs), where memorization, retrieval, and forgetting are architectural operations. This paper asks how such memories can reduce the need for retraining when knowledge changes. For changes expressible as MeMo memory associations, the model's accessible knowledge can be updated by editing explicit memories rather than retraining the whole model. We propose a version-aware operation layer in which high-level operations such as replace, obsolete, keep-history, rollback, and trace are compiled into MeMo-native primitive calls over sequences and tokens. The key observation is that a version-aware operation is rarely a single MeMo association. It is an ordered transaction of primitive edits, for example forgetting one sequence-token chain, memorizing another, preserving a historical chain, and recording an inverse program. The framework introduces two auxiliary CMMs: a Version CMM (V-CMM) for mapping version transitions to transaction handles, and a Transaction CMM (T-CMM) for storing reusable change contents and inverse programs. It supports both direct sequence-level edits and structured diff-level inputs, and outlines an evaluation route for update success, rollback, traceability, locality, and transaction reuse.
展开完整摘要收起摘要↓
MeMo proposes language models with explicit multi-layer correlation matrix memories (CMMs), where memorization, retrieval, and forgetting are architectural operations. This paper asks how such memories can reduce the need for retraining when knowledge changes. For changes expressible as MeMo memory associations, the model's accessible knowledge can be updated by editing explicit memories rather than retraining the whole model. We propose a version-aware operation layer in which high-level operations such as replace, obsolete, keep-history, rollback, and trace are compiled into MeMo-native primitive calls over sequences and tokens. The key observation is that a version-aware operation is rarely a single MeMo association. It is an ordered transaction of primitive edits, for example forgetting one sequence-token chain, memorizing another, preserving a historical chain, and recording an inverse program. The framework introduces two auxiliary CMMs: a Version CMM (V-CMM) for mapping version transitions to transaction handles, and a Transaction CMM (T-CMM) for storing reusable change contents and inverse programs. It supports both direct sequence-level edits and structured diff-level inputs, and outlines an evaluation route for update success, rollback, traceability, locality, and transaction reuse.
作者Taekbeom Lee, Youngseok Jang, Jeonghwa Heo, Jeongjun Choi, H. Jin Kim
Embodied reasoning and exploration are increasingly considered crucial abilities for robots operating in complex and unfamiliar environments. To accomplish tasks in such settings, an agent must identify and acquire the information necessary for the task through exploration. We propose ObsGraph, an observation-centric hierarchical scene graph that unifies scene representation, retrieval, and exploration. It retains visual evidence and organizes it into room-view-object layers: rooms provide coarse semantic anchors, views preserve contextual object covisibility, and objects store fine-grained details. On top of this representation, we perform coarse-to-fine hierarchical retrieval under a bounded budget, and crucially use retrieval outcomes to structure the exploration candidate space--activating room-level exploration, view refinement, or frontier exploration--thereby tightly coupling representation, retrieval, and adaptive multi-scale exploration. Experiments across embodied reasoning and exploration benchmarks demonstrate improved success and efficiency, highlighting the benefits of structured scene representation and more targeted information gathering driven by identified evidence gaps.
展开完整摘要收起摘要↓
Embodied reasoning and exploration are increasingly considered crucial abilities for robots operating in complex and unfamiliar environments. To accomplish tasks in such settings, an agent must identify and acquire the information necessary for the task through exploration. We propose ObsGraph, an observation-centric hierarchical scene graph that unifies scene representation, retrieval, and exploration. It retains visual evidence and organizes it into room-view-object layers: rooms provide coarse semantic anchors, views preserve contextual object covisibility, and objects store fine-grained details. On top of this representation, we perform coarse-to-fine hierarchical retrieval under a bounded budget, and crucially use retrieval outcomes to structure the exploration candidate space--activating room-level exploration, view refinement, or frontier exploration--thereby tightly coupling representation, retrieval, and adaptive multi-scale exploration. Experiments across embodied reasoning and exploration benchmarks demonstrate improved success and efficiency, highlighting the benefits of structured scene representation and more targeted information gathering driven by identified evidence gaps.
作者Dario Pasquini, Michal Bazyli, Taras Fedynyshyn, Artem Sorokin
The use of agentic systems to perform offensive security operations has moved from a theoretical possibility to a commoditized capability. However, while the community has focused on creating more and more capable agents, less attention has been allocated to assessing the security of those systems. In this work, we present the first in-depth security analysis of the most widely used agentic systems for offensive security operations. We show that most of these tools share common design flaws that enable an active adversary to exfiltrate API keys, establish persistent footholds, and fully compromise the operator's machine, even when the agent operates inside a sandboxed container. To support our analysis, we introduce a full cyber kill chain for such agentic systems, capturing the progression from initial LLM manipulation to lateral movement, persistence, guardrail bypass, and sandbox escape. Building on our security analysis, we derive a robust architecture for agentic offensive-security tools and propose actionable, broadly applicable design principles that mitigate the disclosed attack paths at the architectural level.
展开完整摘要收起摘要↓
The use of agentic systems to perform offensive security operations has moved from a theoretical possibility to a commoditized capability. However, while the community has focused on creating more and more capable agents, less attention has been allocated to assessing the security of those systems. In this work, we present the first in-depth security analysis of the most widely used agentic systems for offensive security operations. We show that most of these tools share common design flaws that enable an active adversary to exfiltrate API keys, establish persistent footholds, and fully compromise the operator's machine, even when the agent operates inside a sandboxed container. To support our analysis, we introduce a full cyber kill chain for such agentic systems, capturing the progression from initial LLM manipulation to lateral movement, persistence, guardrail bypass, and sandbox escape. Building on our security analysis, we derive a robust architecture for agentic offensive-security tools and propose actionable, broadly applicable design principles that mitigate the disclosed attack paths at the architectural level.
作者Pranjal Mishra, René Zurbrügg, Max Wilder-Smith, Marco Hutter, Marc Pollefeys, Zuria Bauer, Hermann Blum
Deploying robots in unstructured real-world environments needs accurate, interactive models of the objects. Constructing these models at scale remains a critical bottleneck for robotic system integration. We present ArtiTwinSplat, a framework that automatically constructs articulated, photo-realistic digital twins of objects directly from RGB-D videos, requiring no CAD models, simulation assets, or manual annotations. Our method is built on 3D Gaussian Splatting that preserve geometric fidelity and photometric realism, coupled with an unsupervised articulation discovery pipeline that recovers part structure and joint kinematics from observed motion alone. With tracking and optimization stages our method provides stable, queryable digital twins that support real-time rendering, viewpoint control, and interactive manipulation. Unlike prior methods confined to simulation, ArtiTwinSplat operates directly on real-world observations and produces twins that are immediately usable by downstream robot planning and learning systems. This method offers a practical, scalable pathway toward digital twin construction, lowering the integration barrier for articulated object manipulation in embodied AI and human-robot collaboration contexts.
展开完整摘要收起摘要↓
Deploying robots in unstructured real-world environments needs accurate, interactive models of the objects. Constructing these models at scale remains a critical bottleneck for robotic system integration. We present ArtiTwinSplat, a framework that automatically constructs articulated, photo-realistic digital twins of objects directly from RGB-D videos, requiring no CAD models, simulation assets, or manual annotations. Our method is built on 3D Gaussian Splatting that preserve geometric fidelity and photometric realism, coupled with an unsupervised articulation discovery pipeline that recovers part structure and joint kinematics from observed motion alone. With tracking and optimization stages our method provides stable, queryable digital twins that support real-time rendering, viewpoint control, and interactive manipulation. Unlike prior methods confined to simulation, ArtiTwinSplat operates directly on real-world observations and produces twins that are immediately usable by downstream robot planning and learning systems. This method offers a practical, scalable pathway toward digital twin construction, lowering the integration barrier for articulated object manipulation in embodied AI and human-robot collaboration contexts.
The development of trajectory planning strategies for autonomous ground vehicles (AGVs) represents a prevailing research interest within the domain of intelligent transportation systems. This paper introduces a trajectory planning framework tailored for multi-floor scenarios. The framework consists of two main modules: the task planning module and the trajectory planning module. The task planning module involves a strategic selection phase, where a task planning strategy based on generalized voronoi diagrams (GVD) and multi-objective algorithms is proposed to select the floor exits for each floor. The trajectory planning module utilizes optimization-based methods to generate high-quality trajectories, and a warm-started hierarchical planning framework is designed to ensure rapid convergence. Additionally, for handling complex obstacle constraints, a correlation constraint calculation method is designed for reducing obstacle constraints in trajectory planning. Finally, the feasibility and effectiveness of the proposed framework are verified through simulations.
展开完整摘要收起摘要↓
The development of trajectory planning strategies for autonomous ground vehicles (AGVs) represents a prevailing research interest within the domain of intelligent transportation systems. This paper introduces a trajectory planning framework tailored for multi-floor scenarios. The framework consists of two main modules: the task planning module and the trajectory planning module. The task planning module involves a strategic selection phase, where a task planning strategy based on generalized voronoi diagrams (GVD) and multi-objective algorithms is proposed to select the floor exits for each floor. The trajectory planning module utilizes optimization-based methods to generate high-quality trajectories, and a warm-started hierarchical planning framework is designed to ensure rapid convergence. Additionally, for handling complex obstacle constraints, a correlation constraint calculation method is designed for reducing obstacle constraints in trajectory planning. Finally, the feasibility and effectiveness of the proposed framework are verified through simulations.
作者Viet Hoang Pham, Tran Trung Nguyen, Bao Thu Ho, Phuong Tuan Dat, Thi Thu Trang Nguyen
Speaker recognition has advanced rapidly with large-scale training datasets, yet Vietnamese remains under-resourced, with existing corpora limited in scale and acoustic diversity. Most large-scale datasets rely on facial cues to link speech with speaker identities, restricting data collection to recordings where speakers appear on camera. We propose a face-independent dataset construction pipeline and introduce VieSpeaker, a large-scale Vietnamese speaker recognition dataset. Our approach leverages textual metadata and large language model reasoning to infer speaker identities from transcripts and contextual information. VieSpeaker contains approximately 902 hours of speech from 4,715 speakers. Experiments show that models trained on VieSpeaker achieve improved robustness and generalization compared to existing Vietnamese datasets. This work demonstrates the feasibility of face-independent dataset construction and provides a new direction for building large-scale speech resources.
展开完整摘要收起摘要↓
Speaker recognition has advanced rapidly with large-scale training datasets, yet Vietnamese remains under-resourced, with existing corpora limited in scale and acoustic diversity. Most large-scale datasets rely on facial cues to link speech with speaker identities, restricting data collection to recordings where speakers appear on camera. We propose a face-independent dataset construction pipeline and introduce VieSpeaker, a large-scale Vietnamese speaker recognition dataset. Our approach leverages textual metadata and large language model reasoning to infer speaker identities from transcripts and contextual information. VieSpeaker contains approximately 902 hours of speech from 4,715 speakers. Experiments show that models trained on VieSpeaker achieve improved robustness and generalization compared to existing Vietnamese datasets. This work demonstrates the feasibility of face-independent dataset construction and provides a new direction for building large-scale speech resources.
作者Youran Sun, Xingyu Ren, Chugang Yi, Jiaxuan Guo, Kejia Zhang, Jianda Du, Haizhao Yang
Large language models are making research production scalable, shifting the bottleneck from producing artifacts to judging claims. We present \textsc{Agon}, a research orchestrator that validates what can be checked inside the workflow and leaves the remaining judgments to human scientists. \textsc{Agon} is built on six design principles: Prompt Economy, Future-Facing, Minimal Prompts, OmniDisciplinary, Massive Parallelism, and Zero-Code. We ran \textsc{Agon} across domains for 444 iterations of Prompt Economy loops, using only small starting topics and no human-written experimental code. These deployments demonstrate scalability while exposing new classes of failure. We organize these failures into a taxonomy along severity, fixability, visibility, and capability locus. The taxonomy separates failures the loops can see and fix from those that require human judgment. Together, these results show that \textsc{Agon} is pushing research toward a new paradigm: machine scales, human steers.
展开完整摘要收起摘要↓
Large language models are making research production scalable, shifting the bottleneck from producing artifacts to judging claims. We present \textsc{Agon}, a research orchestrator that validates what can be checked inside the workflow and leaves the remaining judgments to human scientists. \textsc{Agon} is built on six design principles: Prompt Economy, Future-Facing, Minimal Prompts, OmniDisciplinary, Massive Parallelism, and Zero-Code. We ran \textsc{Agon} across domains for 444 iterations of Prompt Economy loops, using only small starting topics and no human-written experimental code. These deployments demonstrate scalability while exposing new classes of failure. We organize these failures into a taxonomy along severity, fixability, visibility, and capability locus. The taxonomy separates failures the loops can see and fix from those that require human judgment. Together, these results show that \textsc{Agon} is pushing research toward a new paradigm: machine scales, human steers.
作者Haoxiang Luo, Tianqi Jiang, Ruichen Zhang, Yinqiu Liu, Gang Sun, Hongfang Yu, Abbas Jamalipour, Dong In Kim
With the rapid development of the Low-Altitude Economy (LAE) ecosystem, Low-Altitude Embodied Artificial Intelligence (LAEAI) agents have become the core carriers of autonomous aerial services, thereby enabling dynamic Low-altitude Computility Networks (LACNets) for distributed computing resource sharing. However, resource-constrained LAEAI agents in decentralized LACNets face a fundamental trilemma of autonomy, security, and efficiency. Existing solutions primarily focus on either optimizing computational performance or enhancing security in isolation, failing to address the inherent trade-offs among trust, performance, and overhead in untrusted dynamic environments with malicious agents. To tackle this challenge, this paper proposes SkyChain Intelligence, a holistic framework that synergistically integrates agentic AI, consortium blockchain, and Multi-Agent Deep Reinforcement Learning (MADRL). We design a lightweight blockchain-based decentralized trust management system with a dynamic reputation mechanism and develop a hybrid-action-space MADDPG algorithm that embeds on-chain reputation scores into the reward function to jointly optimize offloading decisions, resource allocation, and drone 3D trajectories. Extensive simulations demonstrate that our framework outperforms state-of-the-art baselines in task completion latency and energy consumption, while achieving a 94.1% task completion rate in the baseline scenario and stable convergence within 300 training episodes. This work provides a viable path for building secure, autonomous, and efficient machine-to-machine computing ecosystems in the low-altitude domain.
展开完整摘要收起摘要↓
With the rapid development of the Low-Altitude Economy (LAE) ecosystem, Low-Altitude Embodied Artificial Intelligence (LAEAI) agents have become the core carriers of autonomous aerial services, thereby enabling dynamic Low-altitude Computility Networks (LACNets) for distributed computing resource sharing. However, resource-constrained LAEAI agents in decentralized LACNets face a fundamental trilemma of autonomy, security, and efficiency. Existing solutions primarily focus on either optimizing computational performance or enhancing security in isolation, failing to address the inherent trade-offs among trust, performance, and overhead in untrusted dynamic environments with malicious agents. To tackle this challenge, this paper proposes SkyChain Intelligence, a holistic framework that synergistically integrates agentic AI, consortium blockchain, and Multi-Agent Deep Reinforcement Learning (MADRL). We design a lightweight blockchain-based decentralized trust management system with a dynamic reputation mechanism and develop a hybrid-action-space MADDPG algorithm that embeds on-chain reputation scores into the reward function to jointly optimize offloading decisions, resource allocation, and drone 3D trajectories. Extensive simulations demonstrate that our framework outperforms state-of-the-art baselines in task completion latency and energy consumption, while achieving a 94.1% task completion rate in the baseline scenario and stable convergence within 300 training episodes. This work provides a viable path for building secure, autonomous, and efficient machine-to-machine computing ecosystems in the low-altitude domain.
作者Kirill Dubovikov, Omar El Mansouri, Hachem Madmoun, Yanda Li, Sandeep Kumar, Aya El Mir, Supriyo Ghosh, Writabrata Bhattacharya, Adrian Garcia-Garcia, Onkar Pandit, Sunil Kumar Sahu, Federico Castanedo, Larry Murray, Martin Takac, Salem Lahlou
Petroleum-engineering search exposes a supervision gap for strong general retrievers: relevant evidence exists in public web text, but domain relevance labels are scarce. To address this gap, we propose PETRA, a large-scale Petroleum Engineering Text for Retrieval Adaptation dataset and pipeline that converts noisy public web data into a curated domain corpus and synthetic supervision for dense retrieval and reranking. PETRA contains 1.36M curated chunks, approximately 2B token equivalents, $\approx$859k, embedding training rows from $\approx$224k anchors, and roughly 400k teacher-scored reranker candidate rows. Its construction combines high-recall energy-domain curation, an energy-domain classifier with 98.4% test accuracy, chunk-grounded query generation, LLM-written hard negatives, and retrieval-mined candidate lists. PETRA improves first-stage in-domain Normalized Discounted Cumulative Gain (nDCG) from 0.703 to 0.763 through score fusion. Reranker adaptation improves the public Earth Science benchmark by 44% relative and a six-task reasoning-intensive panel by 23%. Failed training recipes show that high train-holdout accuracy on synthetic labels does not predict retrieval gains; retrieval-mined data helps only after being repackaged as teacher-scored candidate lists sampled from the inference-time candidate distribution.
展开完整摘要收起摘要↓
Petroleum-engineering search exposes a supervision gap for strong general retrievers: relevant evidence exists in public web text, but domain relevance labels are scarce. To address this gap, we propose PETRA, a large-scale Petroleum Engineering Text for Retrieval Adaptation dataset and pipeline that converts noisy public web data into a curated domain corpus and synthetic supervision for dense retrieval and reranking. PETRA contains 1.36M curated chunks, approximately 2B token equivalents, $\approx$859k, embedding training rows from $\approx$224k anchors, and roughly 400k teacher-scored reranker candidate rows. Its construction combines high-recall energy-domain curation, an energy-domain classifier with 98.4% test accuracy, chunk-grounded query generation, LLM-written hard negatives, and retrieval-mined candidate lists. PETRA improves first-stage in-domain Normalized Discounted Cumulative Gain (nDCG) from 0.703 to 0.763 through score fusion. Reranker adaptation improves the public Earth Science benchmark by 44% relative and a six-task reasoning-intensive panel by 23%. Failed training recipes show that high train-holdout accuracy on synthetic labels does not predict retrieval gains; retrieval-mined data helps only after being repackaged as teacher-scored candidate lists sampled from the inference-time candidate distribution.
作者Shiyu Li, Ziqi Yan, Zhihao Wu, Jielong Lu, Weiran Liao, Jiajun Yu, Genjie Li, Zeyu Chu, Jiajun Bu, Haishuai Wang
Birth defects are a major cause of fetal loss, neonatal morbidity and long-term disability. In the subset with suspected genetic etiologies, exome and genome sequencing have moved many cases from variant detection to post-sequencing interpretation: clinicians must rank patient-specific candidate variants under incomplete fetal or infant phenotypes and heterogeneous evidence from population genetics, variant-effect prediction, gene-disease validity, phenotype ontologies, cellular and pathway context, protein structure and clinical literature. We present DeepBD, a grounded agentic workflow for variant prioritization and diagnostic interpretation of genetic birth defects. DeepBD organizes the workflow into LLM-assisted case structuring, a pretrained evidence engine, specialist evidence modules and a grounded diagnostic review layer. The evidence engine learns patient-specific variant scores from structured rule evidence, sequence and variant-effect representations and phenotype-conditioned biological context, whereas specialist modules and the agentic layer provide tool-based refinement, candidate-pool review and diagnosis-oriented synthesis from ranked candidates. Developed using an in-house fetal and infant cohort comprising 18,622 cases, DeepBD achieved Recall@1/3/5/10 of 0.658/0.882/0.912/0.929 on an internal held-out solved-case benchmark, outperforming standalone Exomiser, DeepRare and prompted LLM reranking baselines evaluated on Exomiser-derived top-20 candidate variants. Ablation and overlap analyses show that rule evidence, mechanistic context, and specialist refinement provide complementary signals. These findings support a grounded agentic workflow that separates evidence integration, tool-based refinement, and LLM-assisted diagnostic review for retrospective variant prioritization in genetic birth defects.
展开完整摘要收起摘要↓
Birth defects are a major cause of fetal loss, neonatal morbidity and long-term disability. In the subset with suspected genetic etiologies, exome and genome sequencing have moved many cases from variant detection to post-sequencing interpretation: clinicians must rank patient-specific candidate variants under incomplete fetal or infant phenotypes and heterogeneous evidence from population genetics, variant-effect prediction, gene-disease validity, phenotype ontologies, cellular and pathway context, protein structure and clinical literature. We present DeepBD, a grounded agentic workflow for variant prioritization and diagnostic interpretation of genetic birth defects. DeepBD organizes the workflow into LLM-assisted case structuring, a pretrained evidence engine, specialist evidence modules and a grounded diagnostic review layer. The evidence engine learns patient-specific variant scores from structured rule evidence, sequence and variant-effect representations and phenotype-conditioned biological context, whereas specialist modules and the agentic layer provide tool-based refinement, candidate-pool review and diagnosis-oriented synthesis from ranked candidates. Developed using an in-house fetal and infant cohort comprising 18,622 cases, DeepBD achieved Recall@1/3/5/10 of 0.658/0.882/0.912/0.929 on an internal held-out solved-case benchmark, outperforming standalone Exomiser, DeepRare and prompted LLM reranking baselines evaluated on Exomiser-derived top-20 candidate variants. Ablation and overlap analyses show that rule evidence, mechanistic context, and specialist refinement provide complementary signals. These findings support a grounded agentic workflow that separates evidence integration, tool-based refinement, and LLM-assisted diagnostic review for retrospective variant prioritization in genetic birth defects.
Vision-Language-Action (VLA) models are emerging robotic control systems that integrate perception, language understanding, and action generation in a unified architecture. Existing testing approaches for VLA-enabled robots rely on manually constructed symbolic test oracles that determine task success from final environment states. These oracles are costly to construct, require domain expertise, and are often tightly coupled to specific tasks and environments, limiting scalability and reuse. Furthermore, they provide only end-state assessments of task outcomes, offering limited insight into intermediate behavior and fault localization. To address these limitations, we introduce MANGO, a multi-agent framework that automatically generates fine-grained oracles from natural-language descriptions of robotic tasks. MANGO first generates a reusable library of atomic tasks, then generates simulator-grounded oracle definitions for each atomic task, and finally produces executable fine-grained oracles by decomposing complex instructions into ordered sequences of atomic actions and corresponding oracles. The framework uses collaborative Generator, Assessor, and Judge agents that iteratively refine generated artifacts through structured feedback. We evaluate MANGO on the LIBERO_10 and RoboCasa Humanoid Tabletop benchmarks. Results show that MANGO generates executable, fine-grained oracles that detect a similar number of failures as symbolic oracles while accurately localizing them and providing richer diagnostic information. Through ablation studies, we further analyzed component contributions and the effect of initial task set, while preserving oracle quality. Overall, the results show the feasibility and effectiveness of test oracle generation for VLA-enabled robots testing.
展开完整摘要收起摘要↓
Vision-Language-Action (VLA) models are emerging robotic control systems that integrate perception, language understanding, and action generation in a unified architecture. Existing testing approaches for VLA-enabled robots rely on manually constructed symbolic test oracles that determine task success from final environment states. These oracles are costly to construct, require domain expertise, and are often tightly coupled to specific tasks and environments, limiting scalability and reuse. Furthermore, they provide only end-state assessments of task outcomes, offering limited insight into intermediate behavior and fault localization. To address these limitations, we introduce MANGO, a multi-agent framework that automatically generates fine-grained oracles from natural-language descriptions of robotic tasks. MANGO first generates a reusable library of atomic tasks, then generates simulator-grounded oracle definitions for each atomic task, and finally produces executable fine-grained oracles by decomposing complex instructions into ordered sequences of atomic actions and corresponding oracles. The framework uses collaborative Generator, Assessor, and Judge agents that iteratively refine generated artifacts through structured feedback. We evaluate MANGO on the LIBERO_10 and RoboCasa Humanoid Tabletop benchmarks. Results show that MANGO generates executable, fine-grained oracles that detect a similar number of failures as symbolic oracles while accurately localizing them and providing richer diagnostic information. Through ablation studies, we further analyzed component contributions and the effect of initial task set, while preserving oracle quality. Overall, the results show the feasibility and effectiveness of test oracle generation for VLA-enabled robots testing.
作者Ahnaf Atef Choudhury, Md. Parvej Hoque Palash, Shahriar Siddique Ayon, Ramkrishna Saha, Abdullah Al Mamun
One of the significant mental health issues affecting female sex workers (FSWs) is mental disorders, especially depression. Exposure to violence, stigma, and economic hardship further increases their psychological risk. Current machine learning (ML) models are typically ineffective at capturing the high-dimensional and complex risk patterns that exist in this marginalized group. This paper suggests a hybrid predictive model that merges an ensemble feature selection strategy using ANOVA and mutual information and Harris Hawks optimization-tuned logistic regression and represents a new application of swarm intelligence to predict mental health in vulnerable groups. The explainable AI (XAI) methods can be used to understand the factors of trauma associated with model predictions. When applied to a group of 3,005 FSWs, it can be seen that the proposed model is more effective than traditional classifiers, with an accuracy of 95.78%, an F1 score of 95.77%, and an AUC of 0.96, and identifying post-traumatic stress, client-related violence, and occupational factors as major contributors to depression. This work bridges the gaps between conventional and ML approaches to develop an XAI tool that enables vulnerable groups to receive early assistance, evidence-based targeted psychosocial care, and health planning.
展开完整摘要收起摘要↓
One of the significant mental health issues affecting female sex workers (FSWs) is mental disorders, especially depression. Exposure to violence, stigma, and economic hardship further increases their psychological risk. Current machine learning (ML) models are typically ineffective at capturing the high-dimensional and complex risk patterns that exist in this marginalized group. This paper suggests a hybrid predictive model that merges an ensemble feature selection strategy using ANOVA and mutual information and Harris Hawks optimization-tuned logistic regression and represents a new application of swarm intelligence to predict mental health in vulnerable groups. The explainable AI (XAI) methods can be used to understand the factors of trauma associated with model predictions. When applied to a group of 3,005 FSWs, it can be seen that the proposed model is more effective than traditional classifiers, with an accuracy of 95.78%, an F1 score of 95.77%, and an AUC of 0.96, and identifying post-traumatic stress, client-related violence, and occupational factors as major contributors to depression. This work bridges the gaps between conventional and ML approaches to develop an XAI tool that enables vulnerable groups to receive early assistance, evidence-based targeted psychosocial care, and health planning.
作者Saba A. Farahani, Hung Cao, Ramesh Jain, Amir M. Rahmani
Large language models (LLMs) can produce clinically fluent recommendations for type 2 diabetes while failing to satisfy guideline constraints or explicitly justify lifestyle-related glycemic claims. We present T2D-Bench, a reproducible benchmark and evidence-gated evaluation framework for testing whether LLM outputs satisfy explicit, graph-checkable evidence requirements. T2D-Bench is built on a multi-layer clinical-lifestyle knowledge graph that combines a biomedical spine (UMLS, DrugBank, SIDER), computable ADA Standards of Care rules, and lifestyle knowledge connected through a mechanistic bridge to glycemic laboratory effects. Across 100 structured vignettes spanning diagnosis, medication safety, and adversarial lifestyle conflicts, baseline outputs failed benchmark-defined evidence-path checks in 35% of cases for GPT-4o-mini and 33% for GPT-4o. The evidence gate detects unsupported omissions and uses constrained revision to bring outputs into verifier-level compliance with benchmark-defined evidence requirements. These results show that computable evidence constraints can make unsupported clinical omissions explicit, measurable, and correctable in diabetes-focused LLM outputs.
展开完整摘要收起摘要↓
Large language models (LLMs) can produce clinically fluent recommendations for type 2 diabetes while failing to satisfy guideline constraints or explicitly justify lifestyle-related glycemic claims. We present T2D-Bench, a reproducible benchmark and evidence-gated evaluation framework for testing whether LLM outputs satisfy explicit, graph-checkable evidence requirements. T2D-Bench is built on a multi-layer clinical-lifestyle knowledge graph that combines a biomedical spine (UMLS, DrugBank, SIDER), computable ADA Standards of Care rules, and lifestyle knowledge connected through a mechanistic bridge to glycemic laboratory effects. Across 100 structured vignettes spanning diagnosis, medication safety, and adversarial lifestyle conflicts, baseline outputs failed benchmark-defined evidence-path checks in 35% of cases for GPT-4o-mini and 33% for GPT-4o. The evidence gate detects unsupported omissions and uses constrained revision to bring outputs into verifier-level compliance with benchmark-defined evidence requirements. These results show that computable evidence constraints can make unsupported clinical omissions explicit, measurable, and correctable in diabetes-focused LLM outputs.
作者Yucheng Yuan, Yuanfeng Ji, Zhongxiao Li, Ruijiang Li
Spatial proteomics enables single-cell-resolution characterization of protein expression within tissue architecture, playing a critical role in understanding tumor microenvironments and guiding precision medicine. However, current analysis workflows remain fragmented, requiring expert manual orchestration of heterogeneous tools and limiting research scalability and reproducibility. We present SP-Mind, the first autonomous AI agent designed to unify the spatial proteomics analysis pipeline, from raw multiplexed tissue imaging to downstream phenotype discovery. Equipped with expert-curated biological analysis skills and specialized computational tools, SP-Mind converts natural-language queries into end-to-end analytical workflows without task-specific fine-tuning. To rigorously evaluate its capabilities, we introduce SP-Bench, a comprehensive benchmark spanning diverse tissue types, comprising 102 tasks across 18 distinct categories. Through extensive evaluation on SP-Bench and established downstream tasks, SP-Mind achieves state-of-the-art performance compared to existing open-source biomedical agent baselines.
展开完整摘要收起摘要↓
Spatial proteomics enables single-cell-resolution characterization of protein expression within tissue architecture, playing a critical role in understanding tumor microenvironments and guiding precision medicine. However, current analysis workflows remain fragmented, requiring expert manual orchestration of heterogeneous tools and limiting research scalability and reproducibility. We present SP-Mind, the first autonomous AI agent designed to unify the spatial proteomics analysis pipeline, from raw multiplexed tissue imaging to downstream phenotype discovery. Equipped with expert-curated biological analysis skills and specialized computational tools, SP-Mind converts natural-language queries into end-to-end analytical workflows without task-specific fine-tuning. To rigorously evaluate its capabilities, we introduce SP-Bench, a comprehensive benchmark spanning diverse tissue types, comprising 102 tasks across 18 distinct categories. Through extensive evaluation on SP-Bench and established downstream tasks, SP-Mind achieves state-of-the-art performance compared to existing open-source biomedical agent baselines.
We introduce Age of LLM, a turn-based 1v1 benchmark in which two LLMs face off on a 13x7 grid to destroy the enemy base. Three stressors are deliberate: fog of war, full diplomacy (messages, ceasefires, ultimatums; uranium kept secret), and a reliability dimension where every turn must follow a strict JSON schema and an illegal action is silently discarded. The engine is private and each match uses a fresh random map seed and opponent, mitigating the data contamination that affects public benchmarks. Models receive a (near) rule-only prompt with no build-order advice (two tactical seed phrases were present during data collection; see Section 2.7). We benchmark 15 reasoning models across 54 matches and 5,258 actions. Findings: (1) the nuclear rush dominates (78% on the rules-coherent v0.11+ sub-corpus; 85% corpus-wide) with a sole-launcher signature that is largely mechanical under secret-simultaneous launch rules, not a cognitive deterrence failure; (2) military conquest is rare but faster (12.3 vs 18.9 turns); (3) diplomacy is prolific yet almost never consummated; (4) ~58% of illegal actions are fog/state errors, making the illegal-action rate a measure of belief-tracking; (5) — the least established, and the only one we label exploratory — a weak link associates reliability with winning. The corpus is small, unbalanced and not side-swapped, so the ranking is a preliminary descriptive view, not a contribution. Beyond ranking, the turn-by-turn traces of actions and messages make the corpus a lens on how LLMs reason under adversarial uncertainty — their belief-tracking, spontaneous deception, and per-model cognitive "personas" — which we frame as a future research direction. We release the replay format, an isometric viewer and all replays; engine source on request.
展开完整摘要收起摘要↓
We introduce Age of LLM, a turn-based 1v1 benchmark in which two LLMs face off on a 13x7 grid to destroy the enemy base. Three stressors are deliberate: fog of war, full diplomacy (messages, ceasefires, ultimatums; uranium kept secret), and a reliability dimension where every turn must follow a strict JSON schema and an illegal action is silently discarded. The engine is private and each match uses a fresh random map seed and opponent, mitigating the data contamination that affects public benchmarks. Models receive a (near) rule-only prompt with no build-order advice (two tactical seed phrases were present during data collection; see Section 2.7). We benchmark 15 reasoning models across 54 matches and 5,258 actions. Findings: (1) the nuclear rush dominates (78% on the rules-coherent v0.11+ sub-corpus; 85% corpus-wide) with a sole-launcher signature that is largely mechanical under secret-simultaneous launch rules, not a cognitive deterrence failure; (2) military conquest is rare but faster (12.3 vs 18.9 turns); (3) diplomacy is prolific yet almost never consummated; (4) ~58% of illegal actions are fog/state errors, making the illegal-action rate a measure of belief-tracking; (5) — the least established, and the only one we label exploratory — a weak link associates reliability with winning. The corpus is small, unbalanced and not side-swapped, so the ranking is a preliminary descriptive view, not a contribution. Beyond ranking, the turn-by-turn traces of actions and messages make the corpus a lens on how LLMs reason under adversarial uncertainty — their belief-tracking, spontaneous deception, and per-model cognitive "personas" — which we frame as a future research direction. We release the replay format, an isometric viewer and all replays; engine source on request.
作者Lorenzo Rossi, Bartłomiej Marek, Franziska Boenisch, Adam Dziedzic
Assessing the privacy of large language models (LLMs) presents significant challenges. In particular, most existing methods for auditing differential privacy require the insertion of specially crafted canary data during training, making them impractical for auditing already-trained models without costly retraining. Additionally, dataset inference, which audits whether a suspect dataset was used to train a model, is infeasible without access to a private non-member held-out dataset. Yet, such held-out datasets are often unavailable or difficult to construct for real-world cases since they have to be from the same distribution (IID) as the suspect data. These limitations severely hinder the ability to conduct scalable, post-hoc audits. To enable such audits, this work introduces natural identifiers (NIDs) as a novel solution to the above-mentioned challenges. NIDs are structured random strings, such as cryptographic hashes and shortened URLs, naturally occurring in common LLM training datasets. Their format enables the generation of unlimited additional random strings from the same distribution, which can act as alternative canaries for audits and as same-distribution held-out data for dataset inference. Our evaluation highlights that indeed, using NIDs, we can facilitate post-hoc differential privacy auditing without any retraining and enable dataset inference for any suspect dataset containing NIDs without the need for a private non-member held-out dataset.
展开完整摘要收起摘要↓
Assessing the privacy of large language models (LLMs) presents significant challenges. In particular, most existing methods for auditing differential privacy require the insertion of specially crafted canary data during training, making them impractical for auditing already-trained models without costly retraining. Additionally, dataset inference, which audits whether a suspect dataset was used to train a model, is infeasible without access to a private non-member held-out dataset. Yet, such held-out datasets are often unavailable or difficult to construct for real-world cases since they have to be from the same distribution (IID) as the suspect data. These limitations severely hinder the ability to conduct scalable, post-hoc audits. To enable such audits, this work introduces natural identifiers (NIDs) as a novel solution to the above-mentioned challenges. NIDs are structured random strings, such as cryptographic hashes and shortened URLs, naturally occurring in common LLM training datasets. Their format enables the generation of unlimited additional random strings from the same distribution, which can act as alternative canaries for audits and as same-distribution held-out data for dataset inference. Our evaluation highlights that indeed, using NIDs, we can facilitate post-hoc differential privacy auditing without any retraining and enable dataset inference for any suspect dataset containing NIDs without the need for a private non-member held-out dataset.
Large language models are increasingly deployed as agents that reason over documents rather than answer from parametric knowledge. We study archive-grounded reasoning: locating sparse evidence across a large, messy collection of workplace files, reconciling inconsistent terminology, units, and time conventions, and computing an answer. Existing benchmarks address only parts of this setting and none jointly stresses archive-groundedness, agentic exploration, and cross-domain coverage. We introduce Agora, a benchmark pairing 362 questions with eight domain collections of 9,664 authentic documents and 372M tokens, far exceeding any model's context window, so agents must explore deliberately rather than scan exhaustively. Agora is built by an agentic pipeline combining cross-document task synthesis, leakage-preventing obfuscation, and difficulty filtering. Evaluating eight models, we find the task far from solved: even the strongest reaches only 59.4% accuracy, with notable variation across domains.
展开完整摘要收起摘要↓
Large language models are increasingly deployed as agents that reason over documents rather than answer from parametric knowledge. We study archive-grounded reasoning: locating sparse evidence across a large, messy collection of workplace files, reconciling inconsistent terminology, units, and time conventions, and computing an answer. Existing benchmarks address only parts of this setting and none jointly stresses archive-groundedness, agentic exploration, and cross-domain coverage. We introduce Agora, a benchmark pairing 362 questions with eight domain collections of 9,664 authentic documents and 372M tokens, far exceeding any model's context window, so agents must explore deliberately rather than scan exhaustively. Agora is built by an agentic pipeline combining cross-document task synthesis, leakage-preventing obfuscation, and difficulty filtering. Evaluating eight models, we find the task far from solved: even the strongest reaches only 59.4% accuracy, with notable variation across domains.
作者Federico Marcuzzi, Xuefei Ning, Roy Schwartz, Iryna Gurevych
As Large Language Models are increasingly deployed in critical applications, robustly evaluating their social biases is paramount. However, the current literature suffers from widespread methodological fragmentation, which yields contradictory conclusions. This stems largely from ignoring the structural framing of benchmark-level evaluations. To resolve this, we introduce a unified and controllable framework that standardizes heterogeneous benchmarks to systematically contrast isolated demographic assessments with forced-choice comparative settings. Crucially, this allows us to disentangle the confounding effects of Chain-of-Thought reasoning, neutral fallback options, and other structural artifacts in social bias evaluations. Our evaluation across multiple model families reveals a massive, systematic paradigm gap: while isolated assessments limit prejudice activation, comparative settings act as aggressive catalysts for latent discrimination, a shift primarily driven by underspecified contexts. Alarmingly, CoT reasoning exacerbates social biases under comparative settings, and this systemic bias persists as a deterministic prejudice even when models are provided neutral fallback options or claim to answer randomly. Finally, we demonstrate that this comparative prejudice is a generalized phenomenon that scales positively with model size. Ultimately, we offer a crucial methodological guideline: while researchers must leverage comparative settings to robustly audit hidden biases, practitioners cannot safely rely on comparative deployments in ambiguous real-world tasks.
展开完整摘要收起摘要↓
As Large Language Models are increasingly deployed in critical applications, robustly evaluating their social biases is paramount. However, the current literature suffers from widespread methodological fragmentation, which yields contradictory conclusions. This stems largely from ignoring the structural framing of benchmark-level evaluations. To resolve this, we introduce a unified and controllable framework that standardizes heterogeneous benchmarks to systematically contrast isolated demographic assessments with forced-choice comparative settings. Crucially, this allows us to disentangle the confounding effects of Chain-of-Thought reasoning, neutral fallback options, and other structural artifacts in social bias evaluations. Our evaluation across multiple model families reveals a massive, systematic paradigm gap: while isolated assessments limit prejudice activation, comparative settings act as aggressive catalysts for latent discrimination, a shift primarily driven by underspecified contexts. Alarmingly, CoT reasoning exacerbates social biases under comparative settings, and this systemic bias persists as a deterministic prejudice even when models are provided neutral fallback options or claim to answer randomly. Finally, we demonstrate that this comparative prejudice is a generalized phenomenon that scales positively with model size. Ultimately, we offer a crucial methodological guideline: while researchers must leverage comparative settings to robustly audit hidden biases, practitioners cannot safely rely on comparative deployments in ambiguous real-world tasks.
作者Ayan Antik Khan, Harsh Kohli, Yuekun Yao, Huan Sun, Ziyu Yao
Mechanistic interpretability has made substantial progress in automatically localizing circuits, but explaining what localized components do remains labor-intensive and difficult to standardize. In this work, we study whether language model (LM) agents can assist with this explanation problem once a circuit has already been identified. We introduce AgenticInterpBench, a benchmark for circuit explanation built from 84 semi-synthetic transformer circuits with 163 component-level annotations. We propose HyVE (Hypothesize, Validate, Explain), an agentic explainer that analyzes each component through an iterative loop of observation, hypothesis generation, and causal validation, eventually producing a component-level explanation and a circuit-level task description. Across four LM backbones, HyVE recovers useful component- and task-level explanations, but no backbone is uniformly best. Our analysis shows that strong backbones usually form observation-grounded hypotheses, while failures more often arise later in the validation loop, through incomplete validation plans, code execution errors, or unresolved hypotheses. A case study on an arithmetic circuit in Llama-3-8B shows that the same formulation can extend beyond semi-synthetic benchmarks to naturally trained models. Overall, LM agents are promising circuit explainers, but reliable validation remains the key obstacle.
展开完整摘要收起摘要↓
Mechanistic interpretability has made substantial progress in automatically localizing circuits, but explaining what localized components do remains labor-intensive and difficult to standardize. In this work, we study whether language model (LM) agents can assist with this explanation problem once a circuit has already been identified. We introduce AgenticInterpBench, a benchmark for circuit explanation built from 84 semi-synthetic transformer circuits with 163 component-level annotations. We propose HyVE (Hypothesize, Validate, Explain), an agentic explainer that analyzes each component through an iterative loop of observation, hypothesis generation, and causal validation, eventually producing a component-level explanation and a circuit-level task description. Across four LM backbones, HyVE recovers useful component- and task-level explanations, but no backbone is uniformly best. Our analysis shows that strong backbones usually form observation-grounded hypotheses, while failures more often arise later in the validation loop, through incomplete validation plans, code execution errors, or unresolved hypotheses. A case study on an arithmetic circuit in Llama-3-8B shows that the same formulation can extend beyond semi-synthetic benchmarks to naturally trained models. Overall, LM agents are promising circuit explainers, but reliable validation remains the key obstacle.
Transformer-based language models have become the default substrate for natural language processing and the pace of new releases has made it hard for practitioners to separate durable ideas from the noise of incremental announcements. This review works at two levels. At the level of mechanism, we organise the main transformer families into a working taxonomy, covering encoder-only, decoder-only, encoder-decoder, long-context, permutation-based, and generator-discriminator variants. We then extend the discussion to post-2023 developments that changed the picture in practice: instruction tuning, reinforcement learning from human feedback, direct preference optimisation, mixture-of-experts scaling, retrieval augmentation and the current flagship model families from OpenAI, Anthropic, Google, Meta, Mistral and DeepSeek. At the level of use, we survey deployments across healthcare, finance, legal, education, customer service, creative writing and scientific work. Based on this we link each to the specific capabilities that make a transformer the appropriate tool. The contribution of this paper is a critical assessment that is based on the survey. We compare architectures on four axes that matter to deployment decisions, we quantify the trade-off between parameter count and energy cost. We also discuss how alignment methods, data provenance and benchmark saturation change what it means to call a model "state of the art". The final section lists the research questions that we think deserve more attention.
展开完整摘要收起摘要↓
Transformer-based language models have become the default substrate for natural language processing and the pace of new releases has made it hard for practitioners to separate durable ideas from the noise of incremental announcements. This review works at two levels. At the level of mechanism, we organise the main transformer families into a working taxonomy, covering encoder-only, decoder-only, encoder-decoder, long-context, permutation-based, and generator-discriminator variants. We then extend the discussion to post-2023 developments that changed the picture in practice: instruction tuning, reinforcement learning from human feedback, direct preference optimisation, mixture-of-experts scaling, retrieval augmentation and the current flagship model families from OpenAI, Anthropic, Google, Meta, Mistral and DeepSeek. At the level of use, we survey deployments across healthcare, finance, legal, education, customer service, creative writing and scientific work. Based on this we link each to the specific capabilities that make a transformer the appropriate tool. The contribution of this paper is a critical assessment that is based on the survey. We compare architectures on four axes that matter to deployment decisions, we quantify the trade-off between parameter count and energy cost. We also discuss how alignment methods, data provenance and benchmark saturation change what it means to call a model "state of the art". The final section lists the research questions that we think deserve more attention.