Papers

Reinforcement learning for robot manipulation is often bottlenecked by reward design, especially in long-horizon tasks: sparse success rewards provide weak supervision, while hand-crafted dense rewards are tedious to design and generalize poorly across tasks. Progress-based reward models offer a promising alternative by estimating how far an observation has advanced toward task completion, but existing approaches often require task-specific demonstrations or progress labels, and can assign high rewards to visually plausible but physically incorrect states. We introduce the Reference-Anchored Reward Model (RARM), a lightweight visual comparator that converts a single successful demonstration into a dense, progress-aware reward. RARM is trained once on general-purpose videos with a contrastive temporal objective, requiring no robot-specific data, task-specific reward labels, or per-task reward engineering. At deployment, RARM matches rollout clips to reference clips and rewards only confident forward progress, suppressing uncertain matches that may otherwise produce false-positive rewards. Across 9 simulated manipulation tasks from LIBERO and MetaWorld and 4 real-world tasks, RARM achieves the best overall success rates in subsequent RL training, with particularly large gains on long-horizon tasks such as cloth folding, where unreliable progress estimates are especially harmful.

Machine learning systems deployed in dynamic environments frequently operate under nonstationary data distributions, where controlled distribution shift can progressively degrade predictive performance. However, many widely used tabular benchmark datasets lack explicit temporal structure, limiting reproducible evaluation of drift adaptation methods. This work proposes a cluster-induced distribution shift simulation framework that transforms static tabular datasets into controlled evolving data streams through structured perturbations across featurespace partitions. Using this framework, six adaptation strategies are systematically evaluated: static learning, sliding-window retraining, global ADWIN retraining, cluster-local ADWIN retraining, random subspace drift detection, and feature-partitioned drift detection. Experiments are conducted on five benchmark datasets covering both classification and regression tasks using diverse predictive model families, including linear models, k-Nearest Neighbours, tree ensembles, boosting methods, and adaptive online learners.

Subliminal learning lets a student inherit a teacher's hidden trait from distillation data that never names it. We ask when such transfer can be audited before training. The answer is not model identity or scale alone, but channel location: the carrier through which the trait reaches the student. We find three regimes. In a controlled initialization-dependent body channel, a pre-training screen works. Coverage, the cosine between the student's initial distillation update and the teacher's fine-tuning displacement, predicts held-out transfer (Spearman $\rho \approx 0.95$; AUROC 0.997). In pretrained language models, masked single-token traits instead ride convergent vocabulary geometry. This channel is initialization-independent, so initialization-alignment screens, including coverage, are not mechanistic; the useful handles are post-hoc detection and targeted mitigation. Even when a single-token named entity is removed from the loss, the student's held-out probability for that entity rises to 0.40 on average ($\sim 2500\times$), and a related semantic class transfers. In an untied-head model, orthogonalizing the trait's output row against entangled neighbours collapses leakage, while equal-size random-subspace edits do not. Thus removing a target string from distillation labels does not remove the corresponding preference: neighbouring tokens can carry it. Finally, conditional behaviours can route through the network body. For sycophancy, with agreement and correction markers masked from the loss, transfer reaches about 0.63 of the teacher's effect, localizes to body computation, and evades four audits across two model families. We scope this as masked transfer of a condition-present policy. Channel location is necessary for deciding which audits can be sound. It is not a deployment-ready screen: an audit used outside its carrier regime can give false assurance.

We introduce CFAgentBench, a reproducible, self-hostable environment and benchmark for autonomous construction-finance agents: a CFO/controller-class agent operating across the real software stack a US construction finance team runs - ERP, project management, email, documents, pay applications, payroll, certified payroll, lien waivers, and bank/treasury portals. It contains 1,014 machine-gradeable task specifications across 8 domains and 77 families, every family grounded in a real source; a self-validated subset of 40 tasks (54 with a project-management extension) is compiled into oracle-validated executable evaluators, the runnable suite reported here. Following WebArena, the benchmark runs on an executable environment rather than static traces: 35 mock applications (31 reconciled to one company book, plus 4 PM platforms) over 9 archetypes, each implementing a uniform self-hostable app contract, so every task is graded by functional correctness - a state diff plus forbidden-side-effect checks plus required-output regexes - with an LLM judge used only for reply quality, never as reward. A distinguishing principle is a money-movement guard: 278 instances embed a payment, payroll, e-signature, or e-filing step where the correct behavior is to stop and stage for human approval, and executing even the correct transaction fails the task. The public split (n=711) is sized for a 95% Wilson half-width of +/-4.1%; a private, contamination-protected split (n=303) is reserved for remote scoring. In a first three-model open-weight sweep (k=5), the strongest agent reaches pass^1 = 0.67 but only pass^5 = 0.38 - losing 43% of its successes when required to repeat them under temperature-0 decoding. The within-model pass^1 to pass^5 collapse and sharp per-domain heterogeneity are clear evidence that single-attempt accuracy overstates deployable construction-finance competence.

Large language models (LLMs) are increasingly piloted in African healthcare contexts, raising concerns about their potential to manipulate users in high-stakes settings. In a randomised experiment, we examined the manipulative capabilities of two publicly available models, ChatGPT 5.2 and DeepSeek V3.2, among Kenyan participants (N = 303). Participants interacted with either a manipulative variant or a non-manipulative variant before making a treatment decision within a hypothetical clinical scenario. The manipulative variant was prompted to covertly steer participants towards an incorrect treatment option while the non-manipulative variant served as the control condition. Manipulation success rates were higher in the manipulative condition (59.5%) than in the control condition (44.0%), with the effect reaching significance (OR = 2.11, 95% CI [1.12, 4.00], p = .021). These findings highlight the need for improved safety infrastructure specifically targeting manipulation, particularly given the integration of AI into healthcare systems across Africa.

Despite its massive player base and complex hidden-information dynamics, Indian Rummy has received no reinforcement learning attention. Existing agents rely on combinatorial search, which is tactically strong but slow at inference. We present IRumAI, the first RL agent for the domain. IRumAI integrates Proximal Policy Optimization (PPO), meld-aware observation encoding, deadwood-driven reward shaping, and a dual-branch convolutional architecture. IRumAI is RL-trained solely against weak heuristics, after a one-time behaviour-cloning warm-start on stronger demonstration data. It generalises to defeat the entire baseline hierarchy, including a 53.9% win rate against the strongest search-based opponent unseen during RL training. Bypassing explicit search, IRumAI requires just 0.33 ms per action, which is over 7,000x faster than the state-of-the-art heuristic. Ablations validate our architectural choices, and linear probing reveals that the network implicitly models the opponent's hidden hand from public interactions.

Large language models (LLMs) exhibit abilities beyond natural language modelling and text generation. Recent advances in their reasoning capabilities have spurred interest in applying LLMs to complex scientific tasks requiring deep domain expertise and sophisticated reasoning. Quantum computing, as a highly specialised field with significant knowledge barriers and hardware constraints, could greatly benefit from such advancements. However, a key open question that first must be answered is: How can we develop fine-tuning pipelines that instil genuine quantum reasoning in LLMs, rather than task-specific pattern matching? We study this question through quantum circuit simulation as a training objective, where the model must predict the measurement probability distribution resulting from a sequence of quantum gate operations. We propose and compare two fine-tuning pipelines: (1) Supervised Fine-Tuning (SFT) on explicit gate-by-gate state-vector simulation traces, and (2) a two-stage SFT+Group Relative Policy Optimisation (GRPO) approach that sequentially applies SFT followed by GRPO with verifiable rewards. Our findings show that SFT achieves near-perfect in-distribution and gate-count extrapolation accuracy, significantly outperforming both the base model and the GPT-OSS-120B baseline. SFT+GRPO trades some in-distribution precision for better generalisation to larger qubit systems that SFT alone cannot handle. Both pipelines significantly outperform the baselines, demonstrating that targeted fine-tuning on explicit reasoning traces is an effective strategy for advancing quantum reasoning in LLMs.

Physiological signals such as EEG, ECG, and PPG are widely used in clinical monitoring. Recent self-supervised learning (SSL) methods offer an attractive way to leverage unlabeled recordings, yet they still fall short in practice. In particular, current SSL methods struggle across heterogeneous datasets, often distorting clinically meaningful structures or learning shortcuts from temporal and cross-channel redundancy. Consequently, existing SSL methods often deliver limited performance under linear probing, a lightweight adaptation setting that better matches real-world medical scenarios. Moreover, most Transformer-based SSL models encode a flattened spatiotemporal token sequence, incurring high computation and memory cost, and are typically developed within a single modality. To address these limitations, we present SPOTR (Spatio-temporal Pooling One-Token Reconstruction), a compress-reconstruct pretraining framework that introduces a single-token global bottleneck for physiological signals. SPOTR compresses each waveform into a single-token representation and reconstructs the signal conditioned only on this representation. Meanwhile, SPOTR introduces an efficient spatio-temporal compaction module to reduce computation and memory cost. Pretrained on 20 datasets spanning EEG, iEEG, ECG, and PPG, SPOTR consistently outperforms the strongest baseline under linear probing, improving average AUC by 18.49%, 21.71%, 17.86%, and 4.64%, respectively. Compared with a representative general-purpose time-series foundation model, SPOTR achieves around 78% lower latency and 52% lower peak GPU memory on average. The code can be found at https://github.com/5GYYYYY/SPOTR.

We study how the effort and success probability of frontier AI systems scale with human difficulty on problems from Project Euler, an online platform of computational mathematics problems. Our dataset, from the MathArena benchmark, consists of 3840 attempts across 50 problems and 26 model configurations, with problem difficulty measured by the site's public human solve times. Motivated by a proposal of Timothy Gowers, we test a power-law relation $t_{\text{machine}} = a \cdot t_{\text{human}}^b$ between generated-token cost per successful answer and human time, and find $b < 1$ for 20 of the 25 models with usable fits, including the strongest base models; this operationalization therefore does not support an earlier prediction that machines scale worse than humans with difficulty. We also investigate whether success probability on the tested problems can be modeled by a simple exponential decay $p_{\text{success}} = e^{c t_{\text{human}}}$, predicting a linear relation between $\log p_{\text{success}}$ and $t_{\text{human}}$. Using a binning approach for data aggregation we find moderate empirical support (median bin-level $R^2 = 0.92$ across the 22 best-covered configurations) for this model. Following METR, we also fit logistic success curves and extract 50% task-length horizons $h_{50}$; the strongest configurations in our 20 April 2026 snapshot reach roughly $2.5$–$4.3$ hours on our fastest-five human baseline, with a log-linear fit through the state-of-the-art frontier giving a descriptive doubling time of about $75$~days for the SOTA $h_{50}$.

Online planning in continuous partially observable Markov decision processes (POMDPs) using $\omega$-regular specifications requires handling continuous belief dynamics within the finite symbolic memory in order to track temporal progress. Existing methods based on either direct search in belief space or predefined discrete abstractions suffer from drawbacks, e.g., lack of symbolic memory for long-horizon logical progress or difficult to certify from noisy online beliefs. As such, obtaining reliable symbolic states online from continuous observations remains a challenge. To address this issue, we introduce the Revealed Belief Automaton (REBA), an event-driven framework that advances the research from global belief-space discretization to a fundamental new way of thinking, namely online certification of revelation events. Specifically, we propose an online revelation method that, through information-theoretic gates, can dynamically analyse and establish belief abstraction from the continuous belief space by discovering reliable anchors among noisy beliefs. We then develop an incremental topology adaptation mechanism over the certified anchors to realise the online finite Belief Automaton. By combining with the $\omega$-regular specification, REBA is able to support formal parity policy synthesis without a predefined discrete abstraction, which in turn can guide the Monte Carlo Tree Search process to perform online search beyond its local horizon. In addition, we design an error decomposition analysis which can assess the effectiveness and reliability of this discrete guidance for the underlying continuous POMDP. Empirical evaluations in patrolling and navigation scenarios show that REBA matches or exceeds all evaluated baselines, with primary metric gains of +17.0% to +47.4% over state-of-the-art approaches.

Diagnosing mobile crashes in ultra-large-scale industrial applications is a formidable challenge due to the sheer volume of code, the complexity of mixed-language environments, and the inability to reproduce failures locally. Traditional static analysis struggles with scalability, while existing LLM-based agents often rely on reproducible environments unavailable in post-mortem scenarios. We present Holmes, a multi-agent system that automates root cause analysis by synthesizing multimodal runtime signals–stack traces, logs, and thread states–to reconstruct failure contexts without reproduction. Holmes introduces a hierarchical Retrieve-Explore-Reason architecture that leverages low-level artifacts (e.g., registers, assembly) to bridge the semantic gap between open-source business logic and closed-source system frameworks. By dynamically compressing the search space using runtime clues, Holmes precisely navigates 70-million-line codebases to identify non-local defects. Evaluated on real-world crashes from WeChat, Holmes achieves 87.6% accuracy in function-level fault localization and reduces average investigation time by over 98% (to ~77 seconds), demonstrating its effectiveness in transforming labor-intensive debugging into an efficient verification workflow.

Future 6G networks will rely on highly distributed, AI-native Radio Access Networks (RANs), where communication and AI workloads share a common infrastructure. This evolution, combined with increasing deployment density and continuous AI processing, is expected to significantly increase RAN energy consumption. While Open RAN (O-RAN) introduces a programmable and modular control framework through the RAN Intelligent Controller (RIC) and Service Management and Orchestration (SMO), current approaches remain largely policy-driven, limiting adaptive energy-aware coordination across multiple applications. In parallel, AI-RAN promotes the convergence of AI and RAN infrastructures through AI-for-RAN, AI-on-RAN, and AI-and-RAN paradigms, yet efficient mechanisms to jointly orchestrate performance, latency, and energy remain an open challenge. This article proposes an agentic AI-native RAN architecture that bridges O-RAN's structured control with AI-RAN's unified vision. Leveraging semantic intent abstraction and Large Language Model (LLM)-driven coordination, the framework enables adaptive orchestration, conflict resolution, and energy-aware multi-objective optimization across heterogeneous workloads. Through representative AI-for-RAN and AI-on-RAN use cases, we show how such coordination can improve resource efficiency and reduce operational energy consumption, paving the way toward sustainable 6G networks.

Vision Transformers have achieved remarkable success in many fields, yet their deployment on edge devices remains challenging due to their substantial computational demands. Post-Training Quantization (PTQ) offers an attractive solution by compressing models using a small calibration set with minimal training overhead. However, most existing PTQ works adopt a static quantization paradigm that is uniformly applied to all instances. Given the substantial diversity of natural images, the activation distributions vary significantly across samples, making these methods inherently suboptimal. In this paper, we propose ScalePredictor, a dynamic quantization framework for accurate and efficient quantization scale learning of ViTs. We first reveal a hidden correlation between the distribution range of shallow-layer activations and the optimal scales of deeper layers. Based on this, we develop a scale learning mechanism that integrates an efficient range extraction approach to capture robust range statistics at the shallow stage, which are then fed into a Taylor-motivated polynomial scale projection module to generate all quantization scales simultaneously. With the efficiency of polynomial approximation, ScalePredictor introduces insignificant computational overhead while avoiding costly just-in-time calibration. Extensive experiments on ImageNet demonstrate that ScalePredictor consistently outperforms prior PTQ methods, achieving a more favorable accuracy-efficiency trade-off. Code and additional results are shown in the supplementary materials.

Reinforcement learning (RL) has become central to LLM post-training, yet the methods that dominate current pipelines, PPO and GRPO, represent only a narrow slice of what RL offers. Understanding why these methods prevail, and what alternatives exist, requires a principled examination of the design decisions that underlie any RL algorithm. This survey organizes that examination around three stages of algorithm construction. We begin with MDP creation: how the reward function, state space, action space, termination condition, and discount factor are, or could be, defined for LLM training. We then turn to exploration, covering temperature sampling, entropy regularization, intrinsic motivation, tree search, and curriculum learning. Finally, we address learning along four classical RL dimensions: model-free versus model-based, value-based versus policy-based versus actor-critic, on-policy versus off-policy, and credit assignment, including both Monte Carlo methods, which rely on full return estimates, and bootstrapping methods, which update estimates using other learned predictions. Mapping the LLM literature onto this taxonomy reveals a strikingly non-uniform distribution of research effort. Critic-free policy gradients and Monte Carlo credit assignment are densely populated, while value-based methods, off-policy actor-critic training, and bootstrapping-based credit assignment remain largely unexplored despite well-established counterparts in classical RL. These gaps represent concrete opportunities for transferring proven RL techniques to LLM training. By making these gaps explicit alongside the methods that have proven effective, this survey offers researchers in both RL and LLMs a shared framework for understanding current practice and identifying promising directions for future work.

Confidence calibration in large language models (LLMs) is commonly evaluated by comparing predicted confidence with observed accuracy. However, such approaches do not model item difficulty, making it difficult to interpret discrepancies and to determine whether model confidence reflects genuine self-assessment or is merely a byproduct of the response generation process. To address this, we adopt a Rasch model-based latent ability framework and a metacognitive perspective, and propose Latent Confidence Alignment Error (LCAE) to measure the consistency between model self-assessment and the latent error probability implied by model ability and item difficulty. We further incorporate item difficulty as an external signal with a reasoning mechanism. Experiments on a medical-domain dataset with 20 models show that the proposed approach improves self-assessment quality without affecting model ability, and reveals an association between reliability and inference cost.

Cell segmentation is critical for computational pathology and biomedical discovery. While recent Vision Foundation Models (VFMs) have demonstrated remarkable universal feature representations, unlocking their full potential for cellular imaging is currently bottlenecked by resource-intensive adaptation paradigms. Existing methods typically rely on fine-tuning heavy visual encoders, leading to extensive computational overhead and a dependency on large-scale annotations. To address this, we propose the EffiCell-Seg framework for highly efficient cell segmentation without re-training the visual encoder. Our core insight is that pretrained VFMs intrinsically encode complementary structural priors: global saliency for localizing potential cells, and local morphological patterns for delineating cellular structures. To harness these priors, we devise a Cell Structure Prompt Encoder (CSP-Encoder) that synthesizes semantic-aware saliency and principal morphological features from frozen VFM representations into explicit structural prior maps. Moreover, we propose a Synergistic Mask Decoder (SM-Decoder) that enforces contextual consistency by jointly predicting geometric distance fields and semantic maps via mutual cross-guidance. Extensive experiments demonstrate that EffiCell-Seg outperforms state-of-the-art methods across diverse cell imaging modalities while requiring only ~5M trainable parameters, over 130x fewer than fully fine-tuned VFM counterparts. The code is available at https://github.com/xq141839/EffiCell-Seg.

As coding agents are rapidly changing software engineering, a natural question is: what are the core skills needed by future software engineers? To identify where software engineering is headed and thus what skills will be needed, we summarize the results of two round-tables with researchers and industrial practitioners, held in 2026 in New York and Singapore. One key finding is that verification and validation is increasing in importance as agents handle implementation, as highlighted by anecdotes from the events. From our observations, we identify the skills developers need in the agentic era of development, with implications for training and educating future software engineers in coming years.

Scientific discovery can be formulated as an iterative search process over the space of hypotheses and experiments. Contemporary methods navigate this space using heuristics such as MCTS. These algorithms conflate the merit of a hypothesis with the quality of its experimental execution. A promising hypothesis with preliminary execution is therefore ranked below a modest hypothesis whose execution is refined. Moreover, prior methods prune the search logs as the search progresses because the accumulated history outgrows the context window. We propose Agentic Reasoning for Tree Search (ARTS), where we deploy a reasoning language model to navigate this space. The model inspects prior execution logs, diagnoses whether earlier failures arose from faulty implementations or bad hypotheses, and selects the hypothesis to build on next. To mitigate challenges with context length, ARTS uses test-time training to instill the knowledge of search tree in the model weights. Across 22 tasks from MLGym and MLEBench, we show that ARTS outperforms leading algorithms, with over 15.3% relative improvement in the normalized score. With test-time training we show that a Qwen3-4B agent can match performance with closed-source frontier models like Gemini-3 Pro and GPT o3-reasoning with upto 5x lower inference cost. We further observe that on partially observable RL tasks, the test-time trained Qwen3-4B scientist surpasses ARTS with the o3 scientist by rediscovering the human-best recurrent-memory solution that heuristic methods prune away.

In-context learning (ICL) with large language models (LLMs) has emerged as a powerful alternative to fine-tuning for Named Entity Recognition (NER), achieving strong performance with minimal annotation and no additional training. However, prior work has shown that despite their adaptability, LLMs still lag behind fully supervised models such as fine-tuned BERT in structured tasks like NER. While existing studies on ICL for NER have mainly explored few-shot settings, the potential of scaling to hundreds of demonstrations has not been thoroughly investigated. To address this gap, we conduct a comprehensive investigation of many-shot ICL for NER and further explore its effectiveness in annotating and refining data for low-resource NER tasks. Specifically, we evaluate various LLMs across multiple domains using hundreds of ICL examples and then assess the feasibility of using many-shot ICL as a data annotation framework. Our experiments demonstrate that: (1) scaling to hundreds of in-context examples enables LLMs to match or even surpass the performance of fully supervised BERT models; and (2) using about one hundred human-labeled examples as demonstrations, many-shot in-context annotation can generate high-quality labeled data, leading to approximately 10% absolute F1 improvement over existing state-of-the-art approaches when used to fine-tune BERT on low-resource NER.

During hot tests on a production line, engine-sound analysis is crucial to ensuring product quality and performance. However, background noise often interferes with accurate sound analysis, leading to potential errors in engine diagnostics. Traditionally, skilled technicians listen to engine sounds to assess engine health, but this is prone to significant inaccuracies. This study presents an innovative deep learning-based approach to address this issue by removing background noise from engine sound recordings using a U-Net neural network structure enhanced with Residual Attention Blocks (RAB-U-Net). Our intelligent noise removal system significantly improves the accuracy of engine noise detection, outperforming traditional techniques and providing a robust solution for real-time applications in production line environments. This study proposes a novel system for engine noise detection in production lines, marking a valuable advancement for the automotive industry in applying deep learning methods to improve the quality of engine diagnostics.

Conversational AI promises a new kind of preparation for high-stakes workplace negotiations – personalized, interactive, and capable of simulating realistic resistance. That promise is intuitive. We built Trucey, a theory-driven coaching system, to test it. The system encoded four assumptions: that articulation supports clarification, that personalization builds strategic competence, that chunked delivery reduces cognitive load, and that structured scaffolding removes metacognitive burden. A pre-registered experiment (N=267) and interviews (N=15) complicated each of them. Notably, the static handbook we included as a passive control outperformed both AI conditions on empowerment and usability. We reflect on why: each assumption encoded a specific model of how preparation unfolds, and the findings revealed that conversational AI imposes a linear execution model on a task that is fundamentally recursive. We identify an unexamined scope condition on established HAI design guidelines and close with a sequencing principle – map before path, path before simulation – for future AI coaching design.

Foundation models have achieved remarkable performance across medical question answering, imaging, and electronic health record (EHR) tasks, yet reliable clinical deployment remains challenging due to limited interpretability, vulnerability to distribution shift, and weak alignment with clinician reasoning. We argue that these limitations arise because existing approaches prioritize representation learning while treating patient comparison as an emergent property rather than a primary source of clinical evidence. To address this gap, we propose CAFM, a Cohort-Anchored Foundation Model framework that elevates patient cohorts to a first-class object throughout the learning pipeline. The framework consists of four stages: deviation-aware data curation, cohort-conditioned pretraining, multimodal cohort alignment, and clinician-in-the-loop refinement. Together, these stages improve data quality, organize representations around clinically meaningful cohort structure, preserve modality-specific relationships, and support auditable clinical decision-making. The framework is compositional and can augment existing EHR foundation models without modifying their underlying encoders. We illustrate CAFM through four clinical case studies spanning acute kidney injury prediction, cardiovascular risk stratification from electrocardiograms, optic neuropathy triage from orbital imaging, and electroretinogram-grounded report generation. We further present five empirically testable hypotheses and identify open challenges in data quality, irregular temporality, multimodal learning, distribution shift, and evaluation beyond predictive accuracy. We argue that explicitly anchoring foundation models to patient cohorts provides a principled path toward trustworthy clinical AI.

It is tempting to assume any task solvable by a short program can be taught to a model as its chain-of-thought: write the steps out, fine-tune, and the model follows. This paper shows the assumption fails for an identifiable class of procedures. The testbed is nine reasoning tasks, each from a deterministic generator; public and hidden splits share generators, so held-out data proxies test accuracy. I reverse-engineer the generators into Python solvers, render them as chain-of-thought, and distill into a rank-<= 32 LoRA over a 30B (3.5B-active) Nemotron model. Forward-computable tasks install readily: lookup/arithmetic and an 8-bit boolean task transfer (>= 0.99 and 0.68). Cryptarithm does not: distilling its backtracking search holds at 0.01-0.07 across eleven chain-of-thought designs, RL from verifiable rewards, and self-training, even though a search solver answers 71% of instances. This is not a capability gap. The model does the arithmetic on 97-100% of lines and ranks the correct cipher in its top eight on 71%; it cannot carry the search forward as a left-to-right derivation. Fine-tuning learns the shape of a verifiable elimination step while its verdicts become unconditional templates, correct only 16-57% of the time ("verdict-as-token"). The ceiling holds across backbones from 3B to 671B and across fine-tuning and prompting; a controlled intervention isolates the cause: revealing the cipher key, which turns the derivation forward, lifts the same instances from 0.03 to 0.57. When a procedure's only solution is search over information-free structure, no faithful forward chain-of-thought exists to imitate. The task becomes learnable only by removing the search, precomputing its combinatorial core into a catalog and reducing the trace to recall plus verification; the 1st-place solution reaches Private LB 0.92 this way. What distills is memorization and verification, not search.

Streaming text-to-speech synthesis in cascaded LLM-TTS systems still faces latency challenges as most TTS models require full context before initiating generation. We present S5-TTS, a streaming variant of T5-TTS that enables low-latency, word-by-word incremental speech synthesis through encoder-decoder language modeling and monotonic alignment learning. S5-TTS begins generating speech immediately after receiving the first few words, substantially reducing end-to-end response latency. To maintain quality under limited lookahead, we introduce a lookahead-causal masking mechanism with Conv-based auxiliary attention that preserves intelligibility and speaker similarity, and employ interleaved multi-source distillation to further restore naturalness. Experiments show that S5-TTS achieves comparable quality to full-context T5-TTS, supports zero-shot synthesis with high speaker similarity, and significantly reduces end-to-end latency for practical conversational AI systems.

Agentic AI systems retrieve private context, invoke tools, write files, call external services, coordinate with other agents, and may act without human approval. Existing bill of materials artifacts improve transparency for dependencies, model metadata, and training provenance, but leave an agentic transparency gap: capability opacity, the absence of a structured account of what a deployed agent can access, remember, change, delegate, and prove afterward. This paper introduces AgentRiskBOM, a security BOM for risk-scoping tool-using AI agents. It is an additive layer over SBOM, AIBOM, and MLBOM artifacts, referencing them where authoritative while adding fields for runtime authority: autonomy, tool permissions, memory, credential scope, approval gates, audit signals, inter-agent communication, and external action capability. We implement AgentRiskBOM as a JSON-schema artifact with a reproducible corpus, risk scenarios, scorer, diff detector, control mapper, and reports. We evaluate AgentRiskBOM on 13 open-source agents spanning coding, RAG, and multi-agent archetypes, plus 52 risk scenarios across 14 categories. The schema validates all 13 corpus artifacts. Coverage analysis gives AgentRiskBOM a native-equivalent score of 14 across 16 capability dimensions, vs. 1 for SBOM, 1.5 for AIBOM and 2 for MLBOM. Across modeled risk categories, AgentRiskBOM exposes 100% risk-category visibility vs. 10.5% for SBOM-like and 20.9% for AIBOM-like views. To test agentic authority drift, we inject 33 structured deployment mutations; the diff detector identifies the correct change type for all mutations. A secondary penalty-based scorer yields a Spearman correlation of 0.73 with the primary scorer, supporting rank-level consistency while showing that thresholds require human calibration. The results show that agentic AI security needs a machine-readable authority-and-risk artifact before incidents occur.

The Categorical Jacobian of Zhang et al. (2024) reads protein contacts from a language model by perturbing every residue with every alternative amino acid, about $19L$ forward passes. We show the signal it reconstructs is already concentrated in a small subset of attention heads: averaging the top-$K$ contact-relevant heads – selected on as few as 10 labeled proteins, with no fitted per-pair or per-head weights – recovers contacts in a single forward pass and matches or beats the Categorical Jacobian for every bidirectional model where it is defined (bar the smallest, 8M). Our primary test is leakage-clean: on a CAMEO split where neither selection nor evaluation touches data the models have plausibly memorized, the head readout beats the Categorical Jacobian on ESM-2-650M by +9pp ($N = 29$, $p < 0.001$), with the within-model margin reproducing across architectures. Ablations localize the gain to labeled head selection, not to averaging: at a matched label budget the unweighted mean ties a supervised $L_1$ logistic regression on the same heads. Both methods fall 30-36pp from their in-distribution Zhang numbers to the leakage-clean split, which we read as an upper bound on how much prior numbers reflect pretraining overlap. We additionally introduce representation-CJ, a hidden-state generalization of the Jacobian to architectures without a masked-LM head (the output-head-independent analogue of logit-CJ), agreeing with the Categorical Jacobian where both are defined (per-protein Pearson $r \approx 0.95$); show that the optimal $K$ tracks how diffusely a model spreads its contact heads; and find both methods lose the signal on the two causal LMs we test, suggesting attention-encoded pair structure may depend on bidirectional pretraining.

Large language models (LLMs) are increasingly deployed in multilingual settings, yet the energy costs of serving these models across different languages remain poorly understood. We present a systematic study of inference energy consumption across languages with ML.Energy framework (Chung et al., 2026). We find striking disparities: energy consumption per output token varies by up to 8.3 times across languages, while total energy for a fixed set of requests varies by up to 179 times between the cheapest (English, 17.6 kJ) and the most expensive (Pashto, 3,147 kJ) languages. Our analysis shows that this disparity is driven by two compounding factors: (1) higher per-token energy costs for languages using complex or rare scripts, and (2) more tokens generated for low-resource languages. Moreover, we find a double cost + performance penalty: languages with the highest energy footprints also tend to achieve the lowest task accuracy. We reveal that the energy divide persists across models, hardware, and tasks, suggesting a systemic energy inequity in multilingual LLM deployment. Finally, we recommend that the community treat energy as a first-class evaluation axis, extend reporting checklists and model cards to include it, and adopt deployment-side mitigations for better energy efficiency.

Current evaluations of Large Language Models (LLMs) on logical fallacy detection focus on predicted labels, but do not establish whether those labels are supported by the reasoning the models provide. We propose ForEx (Formal Verification for Explainable Reasoning), a framework that translates LLM-generated explanations into Lean4 and verifies whether the translated rationale is derivable under encoded premises, not the logical validity of the original natural language argument. To distinguish prediction outcomes from the formal status of the supporting reasoning, we introduce the LLM Argument Verification Matrix, which separates label consistency from formal verification status. Experiments on LOGIC-Climate show that over 90% of LLM outputs can be translated into formal reasoning chains that pass verification, while agreement with human annotations remains around 20%. These results expose a systematic gap between formal derivability and label agreement, a distinction invisible to prediction-based metrics. ForEx moves LLM evaluation beyond label correctness toward machine-checkable analysis of formalized reasoning chains.

The increasing deployment of large language model (LLM) agents in collaborative workflows demands robust multi-user, multi-principal interaction mechanisms capable of enforcing access permissions, resolving authoritative conflicts, and preventing unauthorized data disclosure. However, a fundamental mismatch exists between the single-user training paradigm of contemporary LLMs and the hard constraints required for multi-principal governance, rendering probabilistic, prompt-based safeguards vulnerable under multi-turn adversarial interactions.Our key insight is that governance constraints – who is authorized, what is restricted, and whose instructions take precedence – are deterministic runtime variables that should be enforced by execution hooks rather than entrusted to the LLM. We present Harness-MU, the first model-agnostic, zero-tuning infrastructure framework for multi-user LLM agents. By decoupling language generation from safety orchestration, Harness-MU guarantees unbreakable permission boundaries while maximizing compliant demand satisfaction. Across four frontier open-weight and proprietary models on the Muses-Bench benchmark, Harness-MU achieves the goal of privacy preservation across all access-control attacks, outperforming the standard baseline by 0.28–0.39 in utility score and improving instruction-following accuracy by up to 48.9 percentage points. Harness-MU advances the philosophy of Harness Engineering, establishing that systematic infrastructure is essential for solving LLM multi-principal governance challenges. The code and data are available at https://github.com/YuanJrShiuan/Harness-MulUser.

Transformer architectures form the foundation of modern natural language processing, yet the Key-Value (KV) cache introduces substantial memory and bandwidth overhead during long-context generation, increasingly bottlenecking large-scale deployment. We propose Keyless Attention, a novel attention mechanism that replaces the conventional key projection with a dedicated value-space routing projection, eliminating key representations from the attention computation entirely and yielding a Value-Only Cache that reduces KV-cache memory by 50% while improving decode throughput. Experiments across multiple models and architectures demonstrate that Keyless Attention achieves comparable perplexity and downstream task performance to standard QKV attention, while consistently reducing KV-cache memory by 50%. Furthermore, Keyless Attention exhibits slower validation loss degradation after the best epoch, indicating improved robustness against overfitting. Ablation studies confirm that the dedicated value-space routing projection is critical, with Keyless Attention outperforming KV-sharing methods that eliminate the key cache without replacing its routing role. Experiments in the pretraining regime further confirm the viability of Keyless Attention in industrial settings.